Road to ruin? Horizontal equalisation of road grant allocations in eastern mainland Australian states.
Drew, Joseph ; Dollery, Brian
INTRODUCTION
Australia has a federal system of government comprised of a
national government, eight state and territory governments, and almost
six hundred local government entities spread over seven local government
systems (McLean 2004). Local government is not recognised in the
Australian Constitution and thus exists as a creature of the relevant
state government 'exercising its [limited] powers by delegation
from the State and under the State's supervision and
authority' (Twomey 2012, 144). The origins of local government
pre-date the establishment of the Australian federation in 1901 with its
genesis in the form of the 1830's Road Boards and Roads Trusts.
This function, together with waste collection and disposal, still form a
significant proportion of the overall functions of Australian local
government (Drew and Dollery 2014a).
Scholars of federalism have attributed numerous benefits to the
federalist model compared with unitary systems (see, for example, Oates
1972). The principal argument resides in 'the hope that state and
local governments, being closer to the people, will be more responsive
to the particular preferences of their constituencies and will be able
to find new and better ways to provide these services' (Oates 1999,
1120). By better aligning public goods and services to local preferences
a greater level of welfare can be attained in the absence of economies
of scale or interjurisdictional externalities (Oates 1999). More
recently arguments in support of federalism have emphasised the
innovative capacity of 'democratic laboratories' (Bednar 2011)
and 'yardstick competition' (Boadway and Tremblay 2012). While
efficiency arguments based on Tiebout mobility have been advanced in
some decentralised systems of government, these arguments have limited
relevance in the Australian local government milieu, given the
restricted range of services provided by municipalities and the
resultant comparatively low property taxes (Drew and Dollery 2014a;
Boadway and Tremblay 2012).
Almost all multi-tier systems of government exhibit a degree of
vertical fiscal imbalance as a consequence of the fact that central
governments typically collect most tax revenue. Since national
governments usually possess greater revenue relative to expenditure
needs than lower tiers of government, including local government, there
is a need for fiscal transfers between the different tiers of
government. In Australian federalism, vertical fiscal imbalance has been
exacerbated by the fact that the federal government has been the sole
collector of income tax since 1942. By contrast, local government is
restricted to a land tax imposed on properties within its jurisdiction,
as well as a range of fees and charges, but even property taxes have
been capped in New South Wales (NSW) (Dollery, Crase and Johnson 2006).
Equalisation grants in Australian federalism have been implemented
'as a necessary counterpart to decentralisation, offsetting its
tendency to create disparities among regions in the ability to provide
public goods or services' (Boadway 2004, 212). The aspiration
underlying fiscal equalisation grants in Australian federalism resides
in the claim that 'a federation with equalised ... fiscal
capacities is one that, in principle, replicates the equity of a unitary
system while at the same time providing the benefits of
decentralisation' (Petchey and Levtchenkova 2004, 192). Such a
system of horizontal fiscal equalisation (HFE) transfers aims to provide
a minimum level of public goods and services to all citizens
irrespective of their spatial domicile (Mieszkowski and Musgrave 1999).
This paper examines the operation of road HFE grants through the
state Local Government Grants Commissions (LGGC) in the context of the
three most populous Australian states. Roads have been selected because
(a) in a large commodity based economy, like Australia, road
infrastructure plays a pivotal role in economic growth; (b) local
government roads comprise about 80% of the total Australian road network
(Chakrabarti, Kodikara, Pardo 2002); (c) road maintenance represents a
quarter of total municipal expenditure and has been cited as a key
factor impinging upon local government financial sustainability
(PriceWaterhouseCoopers (PWC) 2006); (d) road grant funding accounts for
approximately one third of total federal transfers to local government;
and (e) the burden of local road investment and renewal falls especially
heavily on rural and remote councils, which typically possess very low
population densities and large spatial areas.
It is especially important to note that Australian road grant
funding to local government occurs in a financial environment
characterised by harsh fiscal constraints. For instance, in its National
Financial Sustainability Study of Local Government,
PriceWaterhouseCoopers (2006, 111) found that up to 40% of all
Australian local authorities 'could be unsustainable' and the
average annual 'underspend' on 'existing infrastructure
renewals' accounted for between $1.3million and $1.7 million per
municipality. Furthermore, PWC (2006) estimated a national local
infrastructure backlog ranging between $12.0 billion and $15.3 billion
for all local government jurisdictions, with an annual national
shortfall in outlays on existing local infrastructure investment of
between $0.9 billion to $1.2 billion.
The states of NSW, Queensland and Victoria which form the focus of
our enquiry into road grant HFE have been selected on the basis that
they (a) are the most populous jurisdictions in Australia representing a
combined 77% of the nation's population; (b) have all undergone
major structural reform to local government through forced mergers
(Local Government Reform Commission (LGRC) 2007, Independent Local
Government Review Panel (ILGRP) 2013); (c) account for a combined 73% of
total Australian national income (ABS 2013). Furthermore, scholars of
Australian federalism have long been aware of potential inconsistencies
in the distribution of Australian Government transfers by state Local
Government Grant Commissions (LGGCs) (Mathews 1978; Mathews and Jay
1997), as well as the impact this may have on the financial
sustainability of local councils in the different state jurisdictions
(see, for instance, Dollery, Crase and Johnson 2006). However, to date
no attempt has yet been made to quantify the magnitude of the problem of
state LGGCs applying inconsistent methodologies to road funding (Dollery
and Mounter 2010). Accordingly, in this paper we investigate the
magnitudes involved.
The paper itself is divided into eight main parts. Section ii
provides a synoptic account of the HFE literature and Australian
Government legislation on local government transfers. Section iii
outlines the models presently used by the LGGCs in NSW, Victoria and
Queensland. Section iv describes the empirical strategy employed to
assess the consistency of road grant allocations. Sections v and vi
detail the results of re-estimating grant funds using the algorithms of
Queensland and NSW respectively. Section vii presents a comparison of
all three grant allocation methods for the broad strata of urban, rural
and regional councils. The paper ends in section viii with some brief
concluding remarks, including some general public policy implications
for HFE grant systems.
THEORETICAL FOUNDATIONS AND AUSTRALIAN GOVERNMENT LEGISLATION
As we have seen, Australian HFE is complicated by Constitutional
constraints. Since Australian local government systems are
'creatures' of their respective states and territories, there
is no uncontested mechanism to transfer funds from the federal
government directly to municipalities (Dollery, Pape and Byrnes 2008).
In general, this obliges the Australian Government to provide funds
intended for local government to state and territory administrations to
pass on to individual local authorities. However, due to concerns
regarding possible violation of state sovereignty--tested in the High
Court in 1926--the federal government has been loath to impose a uniform
method of allocation onto the various LGGC (Twomey 2012). This
unsatisfactory state of affairs has been the subject of two failed
attempts at Constitutional amendment (1974 and 1988) and an aborted
attempt in 2013 (Twomey 2013). Thus it would appear that local
government will continue to be dependent on LGGC facilitated transfers
into the future. This highlights the need for a long overdue empirical
assessment of the operation of the municipal HFE in Australia.
Petchey and Levtchenkova (2004, 192) have underlined the degree to
which Australian equalisation is motivated by equity concerns. However,
equity between jurisdictions (in this case municipalities) 'is
difficult to comprehend and it carries with it little ethical force in
terms of its policy implementation' (Buchanan 1950, 586). This has
led theorists to assume that 'different persons should be treated
similarly unless they are dissimilar in some relevant respect'
(Pigou 1929, 9). This normative proposition can be interpreted to imply
that individuals who are equally well-off before government policy
intervention should also be equally well-off subsequent to it (Petchey
and Levtchenkova 2004; Mieszkowski and Musgrave 1999). Given the
traditional egalitarian bias in Australian public policy, this approach
has met little resistance (McLean 2004).
Following Buchanan (1950), the equity arguments can be extended at
a municipal level to the proposition (Proposition 1) that if two
individuals are in an equal position in two identical municipalities in
two different states then, for the individuals to remain in equal
positions following public policy intervention, it must be the case that
the two municipalities receive a transfer which is materially the same.
However, in order that 'an individual should have the assurance
that wherever he should desire to reside in the nation, the over-all
fiscal treatment which he receives will be approximately the same'
(Buchanan 1950, 589), a second proposition is implied. This suggests
that the transfers should ensure horizontal equity within states where
broad strata can be readily identified (Proposition 2) (Boadway 2004).
In Australia three local government classifications are
conventionally adopted: urban, rural and regional municipalities. Urban
areas tend to be clustered around the capital cities which in turn are
almost all situated adjacent to the coast. Rural towns cover the bulk of
the Australian continent, typically dominated by agricultural and mining
activity, whereas regional centres are scattered amongst these small
settlements and largely serve as commercial hubs for rural residents.
All three strata have distinct demographic profiles which have been
summarised in Table 1 following the Department of Planning and Community
Development Victoria (2012) classification codes. The existence of three
distinct strata suggests the need to determine whether grant allocation
methods skew transfers to one or the other of the classifications owing
to an exogenous attribute. This represents a subordinate focus of the
paper.
The second approach to horizontal equity delineated above falls in
line with existing Australian Government legislation--the Local
Government (Financial Assistance) Act 1995--which defines HFE as that
which:
'(a) ensures that each local governing body in a State is able
to function, by reasonable effort, at a standard not lower than the
average standard of other local governing bodies in the State; and
(b) takes account of differences in the expenditure required to be
incurred by local governing bodies in the performance of their functions
and in their capacity to raise revenue' (Local Government
(Financial Assistance) Act 1995, s6(3)).
It should be noted that the legislation only provides for potential
ability to deliver local service equality; what actually occurs will be
largely driven by local community preferences and municipal fiscal
effort (Dollery, Kortt and Grant 2013). In fact to achieve actual
service equality 'would violate the very objective of
decentralisation', for actual service equality implies uniform
service levels that would ignore differences in local preferences
(Boadway 2004, 215).
Two additional principles are prescribed in Australian Government
legislation: consistency and transparency. The need for consistency has
been set out in our consideration of the first equity proposition above
(Proposition 1) and it is referred to in the legislation to
'promote consistency in the methods by which grants are allocated
to achieve equitable levels of services by local governing bodies'
(Local Government (Financial Assistance) Act 1995, s3(4)(b)).
In regard to transparency, Duran-Vigneron (2012, 101) has noted
that 'a lack of transparency would undermine the credibility of the
scheme and would be unacceptable from an equity point of view: a
transparent transfer formula defined ex ante ensures that local
governments are subject to the same rule'. The need for a
transparent grants allocation process, in which sufficient information
regarding formulae and data are provided to allow for a full
comprehension of how and why the quantum of transfer is calculated, is
also recognised in Commonwealth statute: 'Increase the transparency
and accountability of the States in respect of the allocation of funds
under this Act to local governing bodies (Local Government (Financial
Assistance) Act 1995, s3(4)(a))
Moreover, the legislation links transparency to accountability:
transparency allows for policymakers, citizens and local authorities to
properly assess the equity of a HFE scheme, but it also acts as a
disincentive to distortion by the political process in the form of rent
extraction (Duran-Vigneron 2012). Accordingly, it represents an
important safeguard to the principle of horizontal equity.
Against this background, the focus of this paper falls on the
analysis of existing road grant allocation methodologies in the eastern
mainland states of NSW, Victoria and Queensland to determine the level
of adherence to the principle of HFE in Commonwealth legislation and the
institutional constraints of consistency and transparency which underpin
its application.
FULL HORIZONTAL EQUALISATION IN FUNDING MODELS
It is useful to provide a synoptic description of the LGGC funding
models currently employed in the three eastern mainland states. The
urban/rural distinction is fundamental to two of the allocation LGGC
methods in response to the infrastructure backlog and fiscal viability
concerns noted earlier for which degree of urban development is a
determinant (see, for instance, Drew, Kortt and Dollery 2014), the key
role that rural road infrastructure plays in the nation's economy
and the higher burden roads place on rural local authorities for road
infrastructure by virtue of population size and density. An example of
the ramifications of these stresses falling especially on rural local
governments can be found in the recent involuntary administration of
Central Darling council in rural NSW which was reportedly no longer able
to pay municipal staff wages. The Council cited inequitable grant
funding allocations as one of the reasons for its insolvency (Brown
2014).
The following descriptions of the three state LGGC methods are
taken from the 2010/11 Grant Commission Annual Reports.
Victorian Road Grant Funding Model
The Victorian Grant Commission Road Funding model is based on
council maintained road length, adjusted for preservation cost and a
series of cost modifiers. Preservation costs are stratified as urban or
rural and calculated according to range of traffic volume (see Table 2).
Cost modifiers include elements for freight loading, sub-grade
conditions, climate, materials and strategic routes. A network cost is
then calculated as the product of the length of roads, asset
preservation cost and overall cost factor. Moreover, an allocation of
$60 per square meter for concrete bridges and $100 per square meter for
timber bridges is added to the network cost. Total road grant funds are
then allocated in proportion to the calculated network costs (VLGGC
2011).
There are a number of problems posed by the Victorian Grant
Commission methodology. Firstly, the assigned preservation costs appear
to confer additional funds for rural roads with traffic volumes up to
1000 vehicles per day, the same funds for volumes of 1000 to 5000
vehicles per day, but lower funds for volumes in excess of this amount.
In fact, rural councils are allocated asset preservation costs of
$4100/km lower than urban councils for vehicle traffic over 5000 per
day. The fact that preservation costs should be higher in rural councils
is consistent with s6(3)(b) of the enabling legislation, which requires
recognition of differing cost structures between councils. However, it
is hard to understand why rural council preservation costs should not
remain higher for the entire range of vehicle traffic. Moreover,
aggregating all the councils as either rural or urban is a crude
approximation of the different cost functions: undoubtedly some local
authorities in remote rural areas would face much higher costs than less
remote rural areas. It also ignores regional centres as the third
important strata.
The second matter regarding the Victorian model relates to cost
modifiers. Whilst it is important to control for varying costs
associated with freight loading, climate, materials, sub-grade
conditions and strategic routes, it does present a few problems. One
problem relates to the transparency requirement embedded in the current
legislation (s3(4)(a)). Whilst the VLGGC does provide index numbers, the
information regarding the calculation of each element is so vague as to
effectively make the VLGGC judgement incontestable. For instance, the
VLGGC report states that 'the raw data for the climate cost
modifier is represented by Thornthwaite Moisture Index Numbers'
(VLGGC, 2011, p 110). However, for actual detail regarding the quite
complex index algorithm one would need to consult the academic
literature, and even then there is a degree of uncertainty given that
the index has been subsequently refined, simplified and otherwise
altered (see, for example, Gentilli 1972). Secondly, only one value for
each cost modifier element is recorded for each council. In the case of
spatially large councils, such as Mildura (2,208,250 ha), it is
difficult to believe that there would not be a number of different
sub-grades, freight loads and climatic conditions. Finally, the climate
modifier relates to moisture only. However, the academic literature has
identified thermal stress as an important predictor of road failure
(Chakrabarti, Kodikara and Pardo 2002).
The final problem posed by the VLGGC approach occurs due to the
allocation of bridge costs directly to the network cost. This implies
that the expenditure associated with bridges is independent of traffic
volume and the various cost modifiers, which is unlikely to be the case.
Although problems undoubtedly exist in relation to the VLGGC
methodology, it can be argued that their approach is superior to that of
the other two states, except for what Duran-Vigneron (2012) argues to be
the most important institutional constraint: transparency. This could be
improved by providing more information and data on exactly how index
numbers are calculated. In addition, adjusting and justifying the
preservation cost rates and adding a traffic volume factor for bridge
costs would also improve the method (other alternatives are explored at
the end of section iii).
Queensland Road Grant Funding Model
The Queensland Local Government Grant Commission (QLGGC) model
draws no distinction between rural and urban councils. It also contains
no cost modifiers for climate, geology, freight or strategic routes. Nor
does the model contain a preservation cost based on the volume of
traffic. Section 2.4 of the QLGGC report (2011) states that the road
funding is allocated simply according to council road length (62.85%
weighting) and local government area population (37.15% weighting).
Population size is based on Australian Bureau of Statistics (ABS) data
(QLGGC 2011, 15).
Perhaps the most commendable aspect of the QLGGC model is its high
level of transparency, consistent with the pressing institutional
constraint embodied in s3(4)(a) of the enabling legislation. However,
three problems are immediately apparent. Firstly, there appears to have
been no effort to abide by the principle of HFE within the state as per
s6(3)(a). In the absence of cost modifiers or preservation cost indexes,
one can only assume that the QLGGC believes that each and every council
faces exactly the same cost structure for road maintenance. Secondly,
the high weighting for population size (37.15%) skews the funding
allocation in favour of urban and regional centre councils. Moreover, it
carries an implicit assumption that traffic volume is somehow closely
correlated to population size. However, no evidence is provided to
support this assumption. Finally, the QLGGC model may produce
inconsistent results given its reliance on ABS population statistics. In
a recent ABS report, intercensal error ranging from 15.2% for
populations less than 2,000 down to 2.4% for populations exceeding
20,000 was identified (ABS 2012). Rural councils are likely to be most
heavily affected by this problem.
NSW Road Grant Funding Model
Appendix 8 of the NSW Local Government Grant Commission report (NSW
LGGC 2011) details its method for allocating federal road grant funds.
The gross NSW road fund allocation is first divided into an urban pool
(27.54%) and a rural pool (72.46%). Urban grant pool funds are then
allocated on the basis of 5% for bridge length. The remaining 95% is
allocated according to road length (60%) and population (40%). Rural
road pool funds are allocated on the basis of 7% for road bridge length.
The remaining 93% is allotted according to road length (80%) and
population (20%).
The NSW LGGC method has a slight degree of HFE within the state
between the crude categories of rural and urban councils, but once again
the distinct strata of regional centre is ignored. However, there is a
great deal of heterogeneity within each of the broad categories, which
cannot be captured by road length and population alone, and this is
likely resolved into a myriad of cost functions, revenue raising
abilities and service standards. The lower weightings assigned to the
rural population size parameter is likely to result in less skewing to
regional centres. However, it will still be present to a certain degree.
The NSW LGGC provides no justification for the actual weightings
assigned, which opens up concerns regarding the transparency of the
process. Finally, the NSW LGGC method also relies on ABS population data
and is thus subject to accuracy problems in intercensal years (ABS
2012).
Focus on the Principles Underwriting Horizontal Fiscal Equity:
Consistency and Transparency
Evans (1991) has observed that 'there is no single correct
principle for grant distribution'. However, the degree of variation
between the three models is astounding in light of the requirement of
the legislation to 'promote consistency in the methods by which
grants are allocated' (Local Government (Financial Assistance) Act
1995, s3(4)). This lack of consistency is critical given that it is a
necessary prerequisite for Proposition 1. Nevertheless, all three models
share common features. Firstly, they all use unaudited council supplied
data which may contain significant errors to varying degrees. Victoria
has, by far, the highest risk associated with such data (via traffic
volume data) followed by Queensland (road length data). Secondly, all
three models fail to account for revenue raising ability, which is an
element of the enabling legislations definition of HFE (Local Government
(Financial Assistance) Act 1995, s6(3)(b). Finally, all three models are
liable to the inaccuracies associated with ABS population intercensal
error--this is greatest for Queensland and NSW, but also applies to
Victoria in as much as the initial interstate allocation is subject to
this problem.
The solution to the problem of horizontal equity being undermined
by a lack of consistency is for either (a) a single method being spelled
out in federal legislation; (b) constitutional reform; or (c) empirical
studies, quantifying the effect on the fiscal position of local
authorities. For reasons discussed in Section ii the first two options
are unlikely to eventuate, which underlines the need for empirical work,
such as that presented in this paper in ensuring the HFE objective is
monitored. With respect to interstate HFE between the three broad
strata, our analysis provides a clear indication of the likely skewing
of transfers towards urban centres in two of the LGGC methodologies. One
way of addressing this violation of Proposition 2 is to allocate road
grant funds in proportion to the moving average of actual road
expenditure as per audited financial statements: this automatically
adjusts for input costs and increased maintenance required as a result
of geology, climate or usage without the use of complex indices (which
may or may not accurately represent reality). A moving average (of
perhaps three years) is advocated in an attempt to thwart gaming by
municipal managers: any attempt to increase the proportion of funding by
inflating road expenditure would have to be sustained over a period of a
number of years to yield results. However, at present consistent
functional reporting of local government expenditure does not
occur--neither within states or between states--and thus accounting
standards would have to be addressed as part of this exercise (Drew and
Dollery 2014b). This obstacle would be best overcome through having the
Australian Accounting Standards Board (an institution established by
federal legislation) amend the relevant standards. Apart from the
horizontal equity advantages of this proposal, it would serve to focus
attention onto road infrastructure: municipalities would be rewarded
proportionally on the effort they put into maintaining infrastructure.
Inefficient practice would still be discouraged by the fact that
councils would only receive a portion of actual costs via the transfer
system. If it were believed that certain strata had higher needs for
funding, then this could be accomplished by an initial division into the
three strata in a manner similar to the NSWLGGC practice. However, it
would be critical to make this division on the basis of empirical data
(for instance, in proportion to non-discretionary spending per capita)
rather than an arbitrary number.
EMPIRICAL STRATEGY
The empirical strategy developed for this paper seeks to answer (a)
whether individuals of equal circumstances but residing in two separate
municipalities (with identical relevant parameters) in two different
states are likely to remain in equal circumstances following public
policy intervention (Proposition 1) and (b) whether HFE occurs within
states according to the three broad strata of urban, rural and regional
(Proposition 2). To investigate the first question we take the cohort of
Victorian councils and re-assess their road grant allocations according
to the Queensland and NSW algorithms. In this way one can assess whether
identical municipalities are treated equally regardless of their state
domicile. To address the second proposition, we first stratify the
Victorian cohort into rural, regional and urban councils according to
the Department of Planning and Community Development Victoria (2012)
classification codes, which are based on the Australian Local Government
Classification codes. This will provide empirical support for our
analysis in Section iii.
The re-estimates were subsequently summarised according to variance
(defined as the real difference between Victorian allocation and the
given comparative model), the percentage variance to Victorian
allocations and the percentage variance to Victorian council operating
results. A positive variance indicated that the figure re-estimated
using the comparative LGGC model was greater than the Victorian
allocation. The percentage variance to operating result has been
included as an indicator of the effect that the alternate models may
have had on a given operating result and, as such, is purely
hypothetical; the statistic simply serves to underline the relative
importance of the inconsistency in grant allocation models (and how this
undermines HFE). Percentage variance to operating result is presented in
absolute terms only, due to the fact that operating results may be
either negative (deficit) or positive (surplus).
Finally, a three state comparison was conducted. This comparison
plotted the results from each of the councils after applying the two
alternative grant funding algorithms. The comparison was also stratified
into urban, rural and regional centres (according to Department of
Planning and Community Development Victoria (2012), classification
codes) in recognition of the specific challenges facing the strata and
our previous discussion of how the existing algorithms may be expected
to skew funding towards centres with high populations.
VICTORIAN COUNCILS RE-ESTIMATED USING QUEENSLAND ALGORITHM
Table 3 presents the results of Victorian council road grants
re-estimated using the QLGGC method. The first row presents overall data
for the state and subsequent rows are the stratifications discussed in
section ii. It will be noted that the mean variance is 0, which tells us
that the two methods both exhaust the entire state road fund allocation,
as they should if our application of the QLGGC algorithm is correct. The
only other remarkable feature of the first row is the size of the
standard errors which suggest an extremely high level of volatility
between the two models.
The positive variance and percentage variance figures for the urban
strata suggest that the Queensland model allocates a higher level of
funding to urban councils than the VLGGC model. This is principally due
to the high weighting given to population in the QLGGC model. As
predicted in section iii, the use of a high population weighting
(37.15%) skews funding towards high population councils, such as those
in urban areas. It thus has the potential to create inequity for
residents in rural and regional municipalities and this is supported by
the evidence presented. The standard deviation is less than the mean and
median suggesting that the trend to higher funding for urban councils
under the QLGGC model has very few exceptions. The percentage variance
is striking in how much difference exists between the two models (on
average approximately 72%). Finally, the percentage variance to
Operating Result suggests that these funding differences could affect
the bottom-line of councils by a significant margin of 5 to 8%.
The statistics derived for the rural strata suggest that, on
average, the QLGGC model allocates fewer funds to rural councils than
the VLGGC model. However, the higher standard deviation indicates that
there are some exceptions to this trend. Once again the percentage
variance is high and suggests little consistency between the models.
Finally, the mean and median percentage variance to operating result
suggests that the variation in grant allocation algorithms is likely to
have a very large effect on rural council bottom lines (approximately 18
to 77%). One of the reasons as to why the variance has such a high
effect on operating results is the smaller revenue base, and thus
relatively higher reliance that rural councils have on grant funding
(Dollery, Garcea and LeSage 2008). Given the financial sustainability
concerns expressed particularly for rural councils (see, for example,
Drew and Dollery 2014; LGRC 2007), this is a finding of major
significance. It suggests that the financial sustainability problems of
Queensland rural councils may be due, at least in part, to the method
for allocating road grant funding.
The relatively high standard deviation for regional centre
statistics suggests a mixed bag. However, the mean and median indicate
that--on average--regional centres receive relatively less road grant
funds under the QLGGC model. This sort of result is expected given that
regional centres tend to have higher population size (mean of 69,572)
than rural areas (mean 19,080), but less than urban (mean 128,487). The
possible effect on operating result is much lower than that of rural
councils, but still quite significant--reinforcing the fact that
consistency in road grant funding is an important issue--because without
consistency it is impossible to argue that allocations are equitable
under Proposition 1.
VICTORIAN COUNCILS RE-ESTIMATED USING NSW ALGORITHM
One of the principal differences between this comparison and that
of section v resides in the almost universally (the exception being
regional centres) lower standard deviations. This suggests much less
volatility between the two methods under consideration here.
For the urban strata the NSWLGGC still allocates higher funds (on
average) than the VLGGC model. However, the size of the variance is much
smaller and this is reflected in the lower percentage variance
(approximately 18 to 19%) and lower percentage variance to operating
result (1.24% to approximately 3%). The same comments essentially apply
to the rural strata which--on average--allocate fewer funds under NSW
LGGC model, but with lower levels of variance (the average variance is
not particularly high but for some small councils, such as Central
Darling, $200,000 annually is doubtless significant). However, the
various measures of variance for the regional centres suggest that the
NSW LGGC model tends to allocate extra funds--on average--when compared
with the VLGGC algorithm.
The relatively lower variance in the N SW LGGC model when compared
to section v is mainly due to the lower weighting assigned to population
under the NSW model (effectively 10.47% for urban and 13.48% for rural
and regional centres when calculated on gross road funds). This
predictably results in less skewing to councils with a higher population
size. The higher effective weighting for the NSW LGGC rural algorithm is
surprising and it appears to be an unintended consequence of the
two-stage distribution of funds.
CONSISTENCY IN ROAD GRANT FUND ALLOCATION--COMPARISON BY COUNCIL
TYPE
Section v and Section vi have compared NSW LGGC and QLGGC
individually with the VLGGC allocations. However, it is still necessary
to compare all three models to one another. Figures 1, 2 and 3 provide
comparative polygon charts for each of the councils represented in the
three strata: urban, rural and regional.
[FIGURE 1 OMITTED]
The results displayed in Figure 1 are unsurprising given the
preceding discussion. In essence, the graphs illustrate that the QLGGC
provides far higher levels of road funding for urban councils than the
NSW LGGC, and the NSW model provides slightly higher levels than the
Victorian algorithm.
When it comes to the rural strata the exact opposite is true. The
Queensland funding model produces the lowest allocations, followed by
the NSW algorithm, then the Victorian model This is perhaps a
significant factor in the fiscal stress widely reported to exist amongst
NSW and Queensland rural councils (see, for instance, the ILGRP 2013 or
LGRC 2007).
[FIGURE 2 OMITTED]
Finally, for regional centres the various road grant fund
allocation models produce a real mixed bag of comparative results.
However, it could be concluded that the NSW model does provide the
highest level of funding for regional centres.
Taken as a whole the empirical estimation presents clear evidence
that any two equally well-off individuals living in two different (but
essentially identical) municipalities located in any two different
states of eastern mainland Australia are not likely to be in equal
positions following federal road transfers. Moreover, the stratified
data confirms that horizontal equity between the broad strata of urban,
rural and regional centres is distorted by algorithms which include
population as a factor.
[FIGURE 3 OMITTED]
LESSONS ON THE IMPLEMENTATION OF HORIZONTAL FISCAL EQUITY TRANSFERS
HFE transfers to Australian local authorities are motivated by
equity concerns. Following Buchanan (1950), the conceptual rationale for
such transfers has been based on the normative egalitarian proposition
that people in like circumstances should remain in like circumstances
following public policy intervention. Due to Constitutional constraints
in Australia, this argument must be stated in two propositions.
Proposition 1 states that if two individuals are in an equal position in
two identical municipalities in two different states, then for the
individuals to remain in equal positions following public policy
intervention, it must be the case that the two municipalities receive a
transfer which is materially the same. The Proposition 2 states that
grant transfer formulae should be constructed in such a manner so as to
ensure there is no skewing towards any of the broad strata on the basis
of exogenous attributes.
Our analysis of the formulae used by the three state LGGCs clearly
showed that the first proposition is routinely violated, despite the
injunctions made in the enabling federal legislation. The analysis of
extant LGGC algorithms also indicated skewing towards centres with high
populations and small road infrastructure burdens in violation of
Proposition 2. This finding was consistent with Twomey's (2012,
180) claim that 'rivers of gold might yet turn to rivers of tears
for local government bodies in the more populous areas if an
equalisation approach to direct funding was taken by the
Commonwealth'. Finally, options to address the failure to achieve
Proposition 1 (more prescriptive legislation, constitutional reform and
greater oversight) and Proposition 2 (the use of the moving average of
actual road expenditure) were canvassed.
We then conducted empirical estimations designed to quantify the
effect of violating HFE principles for the three disparate strata of
urban, rural and regional councils. Consistent with our analysis of the
three models, we found evidence that use of the various methodologies
resulted in very significant differences in the grant transfers, the
greatest of which meant a mean impact on the operating result of rural
councils in the order of 77% (or $457,000 p.a.). For a commodity based
economy with significant road infrastructure backlogs, this result not
only exposes a flaw in Australian Government attempts to provide
horizontal equity, but also provides the first empirical evidence for
why the road infrastructure backlog and municipal fiscal distress in NSW
and Queensland is felt most keenly in rural local authorities (ILGRP
2013, LGRC 2007).
In broader terms, the empirical evidence presented in this paper
can inform the practice of HFE in other local government jurisdictions
in other countries. In the first place, our results provide evidence
against the use of population size as a factor in grant allocation
algorithms if HFE across the broad strata of urban, rural, regional is
to be maintained. Secondly, it was argued that complex indices not only
make review difficult, but also create transparency problems which may
facilitate political controversy and rent extraction. They may also bear
little resemblance to the actual cost pressures facing municipalities. A
better solution might be to use the moving average of audited road
expenditure for councils as the basis for determining the proportion of
transfers to be allocated. This would have the advantage of rewarding
actual effort on road infrastructure and steps could be taken to address
need through an initial division according to each stratum's
proportion of percapita non-discretionary expenditure.
Finally, this study demonstrates that it is not enough to legislate
for the principles underpinning HFE: equity can only be ensured through
regular empirical review and this is largely dependent on transparency.
Moreover, it is unlikely that any static system of transfers can ever
provide horizontal fiscal equity for an indefinite period of time. We
must thus be prepared to adjust practice as new information and methods
come to light.
JOSEPH DREW, BRIAN DOLLERY
University of New England
REFERENCES
Australian Bureau of Statistics (ABS) (2012). Regional Population
Growth, Australia 2011-12. Canberra: ABS.
Australian Bureau of Statistics (ABS) (2013). Australian National
Accounts, 2012-2013. Canberra: ABS.
Department of Regional Australia, Local Government, Arts and Sport
(DRALGAS) (2012). 2009-10 Local Government National Report. DRALGAS:
Canberra.
Bednar, J. (2011). 'The Political Science of Federalism.'
Annual Review of Law and Social Science 7: 269-88.
Boadway, R. (2004). 'The Theory and Practice of
Equalization.' CESifo Economic Studies 50(1): 211-254.
Boadway, R. and J. Tremblay (2012). 'Reassessment of the
Tiebout Model.' Journal of Public Economics 96: 106378.
Brown, E. (2014). Suspended Central Darling Shire Councillors Urge
for Immediate Review of Federal Assistance Grants. ABC: Sydney.
Buchanan, J. (1950). 'Federalism and Fiscal Equity.' The
American Economic Review 40(4): 583-599.
Chakrabarti, S., J. Kodikara and L. Pardo (2002). 'Survey
Results on Stabilisation Methods and Performance of Local Government
Roads in Australia.' Road & Transport Research 11(3): 3-16.
Commonwealth Grants Commission (CGC) (2012). Review Into Improving
the Impact of Financial Assistance Grants on Local Government Financial
Sustainability. Canberra: CGC.
Commonwealth of Australia. Local Government (Financial Assistance)
Act 1995.
Department of Planning and Community Development, Victoria (DRCDV)
(2012). Guide to Local Government. Melbourne: DRCDV.
Dollery, B., L. Crase and A. Johnson (2006). Australian Local
Government Economics. Sydney: University of New South Wales Press.
Dollery, B., M. Kortt and B. Grant (2013). Funding the Future.
Sydney: Federation Press
Dollery, B., J. Garcea and E. LeSage, Jr. (eds) (2008). Local
Government Reform--A Comparative Analysis of Advanced Anglo-American
Countries. Cheltenham: Edward Elgar.
Dollery, B., M. Kortt and B. Grant (2013). Funding the Future:
Financial Sustainability and Infrastructure Finance in Australian Local
Government. Sydney: Federation Press.
Dollery, B. and S. Mounter (2010). 'Local Infrastructure
Investment, Maintenance and Renewal: A Comparative Analysis of
Contemporary Australian and New Zealand Local Government.'
Australasian Journal of Regional Studies 16(2): 217-232.
Dollery, B. E., Pape, B. and J. Byrnes, (2008).
'Constitutional and Federalist Perspectives on the Australian Local
Government Roads to Recovery Funding Program', ICFAI Journal of
Governance and Public Policy, 3(4), 34-51.
Drew, J. and B. Dollery (2014a). 'Keeping it In-House:
Households Versus Population as Alternative Proxies for Local Government
Output.' Australian Journal of Public Administration, In Print.
Drew, J. and B. Dollery (2014b). 'Inconsistent Depreciation
Practice and Public Policymaking: Local Government Reform in New South
Wales.' Australian Accounting Review, In Print.
Drew, J., M. Kortt and B. Dollery (2014). Economies of Scale and
Local Government Expenditure: Evidence From Australia. Administration
& Society In Print.
Duran-Vigneron, P. (2012). 'When Transparency Goes Against
Efficiency: Fiscal Equalization in a Context of Firm and Household
Mobility.' Journal of Public Economic Theory 14(1): 99-130.
Evans, A. (1991). 'A Theoretical Model for Local Road
Grants.' Urban Studies 28(2): 219-231.
Gentilli, J. (1972). Australian Climate Patterns. Melbourne: Thomas
Nelson.
Grants Commission Act 1973. Commonwealth of Australia.
Independent Local Government Review Panel (ILGRP) (2013). Future
Directions for NSW Local Government--Twenty Essential Steps. Sydney:
ILGRP.
Local Government (Financial Assistance) Act 1995. Commonwealth of
Australia.
Local Government (Personal Income Tax Sharing) Act 1973.
Commonwealth of Australia.
Local Government Reform Commission (LGRC) (2007). Report of the
Local Government Reform Commission Volume 1. Brisbane: LGRC.
McLean, I. (2004). 'Fiscal Federalism in Australia.'
Public Administration 82(1): 21-38.
Mathews, R. L. (1978). Australian Federalism, Centre for Research
on Federal Financial Relations, Australian National University.
Mathews, R. L., and W. R. C. Jay (1997). Federal Finance:
Australian Fiscal Federalism from Federation to McMahon. Centre for
Strategic Economic Studies Victoria University.
Mieszowski, P. and R. Musgrave (1999). 'Federalism, Grants and
Fiscal Equalization.' National Tax Journal 52(2): 239-260.
New South Wales Local Government Grants Commission (NSWLGGC)
(2011). NSW Local Government Grants Commission 2011-12 Annual Report.
Sydney: NSWLGGC.
Oates, Wallace E. 1972. Fiscal Federalism. NY: Harcourt Brace
Jovanovich.
Oates, W. (1999). 'An Essay on Fiscal Federalism'.'
Journal of Economic Literature XXXVII September 1999: 11201149.
Petchey, J. and S. Levtchenkova (2004). 'Fiscal Equalisation
in Australia: Proposals for an Efficiency-Based System.' Economic
Papers 23(2): 189-200.
Pigou, A. (1929). A Study in Public Finance. London: Macmillan.
PriceWaterhouseCoopers (PWC) (2006). National Financial
Sustainability Study of Local Government. Sydney: PWC.
Queensland Local Government Grants Commission (QLGGC) (2012).
Queensland Local Government Grants Commission Annual Report 2012.
Brisbane: QLGGC.
Twomey, A. (2012). 'Always the Bridesmaid--Constitutional
Recognition of Local Government.' Monash University Law Review
38(2): 142-180.
Twomey, A. (2013). Local Government Funding and Constitutional
Recognition. Sydney: University of Sydney Constitutional Reform Unit.
Victorian Local Government Grants Commission (VLGGC) (2011).
Victorian Grants Commission Annual Report 2010-11. Melbourne: VLGGC.
Victorian Local Government Grants Commission (VLGGC) (2012).
Victorian Grants Commission Annual Report 2011-12. Melbourne: VLGGC.
Table 1
Statistical Means of Key Descriptors of Municipal
Classifications
Population Population Kerbed Roads Unkerbed
Size Density (km) Roads (km)
(persons/
[km.sup.2])
Urban 128,487 1718 500 184
Rural 19,080 6.6 102 2,222
Regional 69,572 71.2 356 1,893
Centres
Table 2
Road Preservation Costs in Victoria
Road Type Daily Traffic Volume Standard Annual Asset
Range Preservation Cost $/km
Urban <500 $3,600
500-1000 $4,900
1000-5000 $6,600
5000+ $10,700
Rural Natural Surface $350
<100 $2,500
100-500 $5,200
500-1000 $5,800
1000+ $6,600
Source: Victorian Local Government Grants Commission (2012)
Table 3
Victorian Councils Re-estimated Using Queensland Algorithm
Measure of Variance ($) Percent Variance Percent Variance
Central to Victoria to Operating
Tendency Result (absolute
value)
Mean 0.0 21.57 40.80
Median -36,671 -3.3 7.3
Standard 737,319 61.57 194.24
Deviation
Urban
Mean 516,473 72.25 8.36
Median 583,765 72.87 5.23
Standard 458,395 57.54 11.01
Deviation
Rural
Mean -457,103 -18.53 77.36
Median -425,495 -24.88 17.94
Standard 710,169 30.03 284.72
Deviation
Regional
Mean -6,495 -3.32 15.50
Median -107,686 -7.05 4.26
Standard 466,886 21.34 35.96
Deviation
Table 4
Victorian Councils Re-estimated Using NSW Algorithm
Measure of Variance ($) Percent Percent
Central Variance to Variance to
Tendency Victoria Operating
Result
(absolute
value)
Mean 0 7.82 30.21
Median 30,416 2.98 2.96
Standard 538,705 28.22 146.83
Deviation
Urban
Mean 88,775 19.29 3.02
Median 139,179 17.88 1.24
Standard 271,323 23.39 5.75
Deviation
Rural
Mean -199,603 -5.51 58.24
Median -170,060 -9.8 6.16
Standard 640,729 29.21 214.15
Deviation
Regional
Mean 394,993 18.05 17.54
Median 378,368 21.07 2.96
Standard 508,172 18.83 46.68
Deviation
