Equivalent salaries in five Australian capital cities.

Henman, Paul

Introduction

Over the last decade, the spatial or geographical nature of social and economic indices has proliferated. More than ever, we are now armed with knowledge about the areas of socio-economic advantage and disadvantage. Indeed, research suggests that the fissures of socio-economic outcomes have become more pronounced. Moreover, overlaying geographical statistics helps to develop an understanding of the dynamics reinforcing one's socio-economic position. (1)

While these statistical developments are of considerable academic interest, the geographical nature of disadvantage has important public policy implications. For example, the appreciation of the geographical concentration of poverty has led to a whole gamete of programs focused on individual communities. This is perhaps most clearly expressed in the UK's New Deal for Communities (Lawless 2006), but similar Australian examples exist. Such programs focus on disadvantaged communities as a way in which to build local human and social capital in an effort to enhance the prospects of individuals living in those communities. Another policy response to geographical disadvantage has been to reform housing support programs to enable disadvantaged families to re-locate in areas with higher levels of employment. Geographical disadvantage results from a range of factors, with the geographical distribution of living costs one such factor that raises a range of related, but separate policy issues.

Governments have responded to the geographical distribution of living costs by offering tax deductions for residents in high-cost areas. In Australia, the Zone Tax Offset provides tax concessions for people living in remote areas and Remote Area Allowance (Centrelink 2007) is a benefit for income support recipients living in a remote area. In the UK it is the major metropolitan area that accompanies such support in the form of a London Allowance paid by employers.

While constant media coverage ensures that we are well versed in the cities with the highest costs of housing, there is remarkably little research on the differential living costs in Australian capital cities. The Australian Bureau of Statistics (ABS) does publish a short list of the differential costs of major grocery items as part of its Consumer Price Index (CPI) research (ABS Cat. No. 6403.0), but not much else exists. Internationally, the OECD's Purchasing Power Parity indices provide a means to make financial comparisons between countries (OECD 2002).

This paper examines the effects of differential costs in five Australian capital cities to determine the various salaries required in each city to obtain an equivalent standard of living. It answers the question, if a household earns $X in Brisbane, what salary is required to achieve the same standard of living in a different capital city? In answering this question, this paper considers a range of household types; from low to high income, and single and couple headed households. While any city could have been used as the comparator, Brisbane was chosen as it sits in the middle of the spectrum from most expensive to cheapest capital city in terms of housing costs.

The paper begins by explaining the method used to answer the above question. It summarises prior research which has been extended for the purposes of this study. The next section discusses the results, and is followed by a brief conclusion which considers the policy implications of the research.

Method

There has been very little research that calculates equivalent living standards in Australian capital cities, and none that calculates equivalent salaries. However, the issues of regional comparisons have been well discussed in relation to the treatment of housing costs. Siminski and Saunders (2004) have given extended consideration to these and other issues relating to regional comparisons, as discussed below.

A central concern for regional comparisons is identifying a way in which comparisons can be adjusted to account for regional differences that impact on living standards. There are two elements to this: regional variations in prices; and regional variations in needs (such as different energy and clothing requirements resulting from different climates). The approach used in this study is similar to that used internationally to generate Purchasing Power Parity indices. This involves comparing regional prices for an "identical" basket of goods and services required by households to achieve the same standard of living. Pricing an identical basket of goods and services in different locations picks up the first of the two above elements of regional variations in living standards. Slightly varying the basket for such things as clothing and energy to take account of climatic differences ensures that the second element is dealt with. While not identical baskets, they provide identical living standards.

Such an approach is called the budget standards method. It has traditionally been used to derive poverty measures, but increasingly budget standards are calculated for median and even higher living standards and used for a variety of purposes. The research reported here builds on the seminal Australian budget standards research for Sydney households conducted by the Social Policy Centre Research Centre (SRPC) at the University of New South Wales (Saunders et al 1998) and subsequently extended to other capital cities by Henman (2001). (2)

The budget standards methodology provides a relatively straightforward approach for calculating equivalent household expenditure. Identical budget standards--a specific basket of goods and services required by a household to meet a particular standard of living--are costed using local prices in each location (i.e. capital city) under examination. Various living standards can be defined by defining baskets to reflect each living standard. One clear advantage of the budget standards approach is that equivalent living standard levels are based on costing "identical" basket of goods and services in different locations. As such, different expenditure levels in different localities are guaranteed to reflect identical standards of living. Budget standards are also sensitive to other circumstances and requirements of different household types, such as the number of adults and their labour market status, the age and sex of the children, and housing tenure.

The budget standards approach is also more transparent than those that use complex mathematical models derived from expenditure survey data. This is because new living standard levels can be constructed by adding or subtracting items from the basket (or altering their quantity and quality) and re-costing the basket. A key constraint of budget standards is access to appropriate data on relative prices of goods and services. (3)

The budget standard methodology is, however, not readily suited for estimating expenditures required to achieve high and very high living standards. This is because high living standards have high levels of discretionary income, which in turn generate a great variety in the way households choose to spend and save this income. As a result of this variability at high living standards, it becomes problematic to define a specific basket of goods and services at these 'affluent' levels. (4) When equivalent living standards at high living standard levels are required, the budget standards methodology is best combined with behavioural data that provides average estimates of expenditure for high-income households. This hybrid approach is adopted in this paper for calculating a 'luxury' standard of living. A further weakness of the budget standards approach is that the estimates technically relate to quite specific household types and cannot be immediately generalised to a range of households. However, this weakness has minimal impact in this study.

It is worth observing that Henman (2001) derived separate capital city budget standards from the Sydney ones by taking account of regional differences in prices using relativities in child care costs from government administrative data, housing costs from Real Estate Institute data and ABS pricing data for other items. In addition, the energy budgets were adjusted to take account of regional differences in energy needs.

In generating regional budget standards the difficulty in adjusting housing costs must be acknowledged. The central issue is that identical dwellings in different locations to not provide identical living standards, what Castles (1997) calls "comparison resistant goods". This is due to the different locational benefits of different locations. It is typically held that differences in housing costs reflect different consumption benefits associated with each location. For example, higher housing prices in Sydney is interpreted as reflecting better quality dwellings and communities, and better access to public and private infrastructure and services, including transport, retail, employment (and their wage levels), climate and leisure. Such locational advantages and disadvantages are extremely hard to measure, with competing elements and a great level of subjectivity. One person may think medium density a plus, but another would regard it a minus. A study by the Department of Social Security found that after taking account of a range of dwelling type and locational advantage characteristics--including age of dwelling, socio-economic status of area--there remained considerable differences in median rents between capital cities (1996: Figure 5.4). Due to the locational benefits perspective, regional comparisons of living standards have typically removed housing costs from consideration, as is done in calculating Purchasing Power Parity.

Apart from locational benefits, the economic model allows for another driver of price--demand relative to supply. Price inflation increases when demand outstrips supply. While some aspects of demand certainly relate to perceived locational advantage, others relate to other factors such as relative advantages of investment in housing versus shares or other forms of investments (5), insufficient release of land for housing developments, increased immigration and internal migration. Moreoever, given that housing is not mobile, the propensity of people to live within a fixed geographical location means that housing markets do not operate like other goods. In short, while locational advantages may effect housing prices, price differentials are also importantly influenced by demand/ supply pressures.

To reinforce this message, consider the high relative housing costs of Sydney. In early 2007, Sydney's median house price was the highest in the country, some 36 per cent greater than Melbourne and 50 per cent greater than Brisbane (REIA 2007). While ABS does not publish average weekly earnings for capital cities, the differentials in wages relative to New South Wales in the same time period was a much smaller 6 and 15 per cent respectively. According to the locational advantage thesis, there must be some other very big advantages in Sydney. Moreover, Sydney's dwellings are, on average smaller in housing and allotment size, than Brisbane. Melbourne arguably has better public transport than Sydney, which is better than Brisbane's. The climate is undoubtedly best in Brisbane. On these important indictors, there is no clear advantage of Sydney justifying the significant price disparity.

Siminski and Saunders (2004) argue that housing costs involve three components: consumption of dwelling; consumption of locational goods; and locational wages. Ideally, the contribution of each of these components should be considered when making regional comparisons. Due to data limitations this is not possible, but they conclude that housing costs should be given greater consideration in regional analyses. Following Siminski and Saunders (2004) this paper takes the view that housing price differentials are reflective of costs of living differentials, although it is acknowledged that some of the differential is also due to locational benefits (and hence we are not treating like with like). Thus, the final results should be interpreted with this rejoinder in mind.

Equivalent salaries were calculated from Henman's (2001) regional budget standards in a step-wise fashion, as follows. First, up-to-date budget standards at five different levels, and for various household types were calculated for Brisbane, Sydney, Melbourne, Adelaide and Perth. Previous budget standards from Henman (2001; 2007a) were updated using Australian Bureau of Statistics (ABS Cat. No. 6455.0.55.001) data on changes in prices, specifically the CPI detailed items for each capital city. The ten budget standards components--housing, energy, food, clothing, household goods and services, child care, leisure, transport, health and personal care--were updated using the most compatible sub-components of the CPI. (6) Thus, the food and childcare components for the Brisbane budget standard were updated from Henman (2001) using the food and childcare sub-components of the ABS CPI.

Housing costs--which account for 41 per cent of total household costs on average-contribute the greatest variation in budget standards between capital cities. Child care, which ranges from 0 to 23 per cent of total household costs, is also a large contributor to locational variation. Food costs, on average 18 per cent of the household budget, is the third largest factor contributing to regional price variations.

The five different living standard levels used in this study were derived from the original two living standards developed by the SPRC; a "low cost" representing low-income households and a 'modest but adequate' standard representing middle Australia. Using the budget standards from Henman (2001), a "low cost+" living standard was derived by assuming the low cost household was purchasing their own home at the first quintile price in the outer ring of their city, rather than in private rental. The "modest but adequate" living standard was altered by allocating a slightly larger dwelling for households than that specified by Saunders er al (1998). A distinction between "modest but adequate" and "modest but adequate+" was derived whereby the former was in private rental and the latter a purchaser.

Due to the difficulties in constructing budget standards for high living standard levels, a "luxury" standard was created with the aid of behavioural data. Specifically, the ten different components of the 'modest but adequate+' budget standards (eg. energy, food, clothing) were adjusted according to the relativities of these expenditure components between the middle and top quintile households from the Household Expenditure Survey 2003-04 (ABS 2007). The behavioural data took into account different household types. (7)

Table 1 specifies the 34 different household types for which budget standards were calculated. These households are varied by living standard level, household tenure and labour market status.

Three different approaches were taken to measuring housing costs to assess the sensitivity of housing costs to overall outcomes. One method was to assume private rental at current entry level rents (the Low Cost and Modest But Adequate columns in Table 1). A second approach was to assume home purchasers, but at two different time periods. Either purchasing a house in the current quarter or 10-15 years previously in the March quarter 1993. Mortgage repayments were based on current variable interest rates for 85% of the dwelling purchase price.

In order to track the differential between capital cities over time, budget standards were calculated for three different points in time: March quarter 2003; September quarter 2005; and September quarter 2006 (available from the author).

Having identified equivalent expenditure levels using these budget standards, the expenditure differentials between Brisbane and other cities were modelled. Multiple regressions were used to derive an equivalence function using the five points of equivalent expenditure in capital cities. A log-linear regression equation was used according to the following formula. (8)

[E.sub.s] = a [E.sub.B] + b ln([E.sub.B]) + c L + d H + e (1)

where the input variables are:

* Brisbane expenditure ([E.sub.B]);

* Housing type (H) (0 = private renter; 1 = owner occupier);

* Labour market status of second adult (L) (0 = not in the labour force; 1 = employed full-time);

* Equivalent other city expenditure ([E.sub.S], [E.sub.A], [E.sub.M] [E.sub.P]).

Having derived equivalence expenditure equations--whereby given an expenditure level in Brisbane the expenditure required in another city to obtain the same living standard--equivalent expenditure were translated into gross income points using an Excel-based tax-benefits model developed by the author. (9) With a given private income in Brisbane, this tax-benefits model calculated a household's disposable income, which was then taken as the expenditure level inserted into the equivalent expenditure functions. The resulting equivalent expenditure in a different capital city was then converted into a household's private income (ie before tax and benefits) using the tax-benefit calculator in reverse.

Findings

For the March quarter 2003, equivalent salaries in other cities were calculated for Brisbane households with salaries of $35,000, $50,000, $75,000 and $100,000 and for five different household characteristics (single adult; couple adult; couple with one child; couple with two children; sole parent with two children). For each household type, variations in housing tenure (private rental or purchasing) and in labour force status of the second adult were also calculated (see e.g. Table 2). In households with more than one working adult, a range of household incomes was used derived by combining salaries from the above levels to take account of the impact of the labour force status of the second adult (and thus associated child care costs) and the consequent tax-social security policy mix. Wage relativities between husband and wife were maintained to estimate relative wages in other cities. Incomes for the later two time periods were inflated by the seasonally adjusted wage price index (ABS Cat. No. 6345.0).

Table 2 presents the results of these calculations for the couple with no child household types for September quarter 2006. (10) The averages for Tables 3 to 5 have been obtained by using a simple mean of the different household types. These statistics therefore are averages of the range of household types defined by the budget standards and are not measures of disparities relating to the Australian population. Table 2 illustrates the range of household private income levels for which equivalent salaries were calculated. The Table should be read across each row. Thus, for example, the Table demonstrates that the equivalent salaries for the Brisbane private renter couple with no children having salaries of $58,000 per annum and $29,000 per annum are: $62,700 and $31,300 in Sydney; $59,800 and $29,900 in Melbourne; $52,700 and $26,300 in Adelaide; and $60,100 and $30,100 in Perth.

The overall results across all household types indicate that in the September quarter 2006, equivalent salaries relative to Brisbane are higher in Sydney, mostly higher in Melbourne (although to a lesser extent), almost always lower in Adelaide, and sometimes lower and sometimes higher in Perth.

The results also show that equivalent salaries between Australian capital cities can vary considerably in dollar amounts. As expected, as salaries in Brisbane increases, the variation generally increases. As a result, for some household types, earning at least $100,000 in Brisbane requires an extra 50 per cent more private income in Sydney to achieve the same standard of living. This is a result of two main factors: the greater housing costs in Sydney, particularly for mortgagees; and the high marginal taxation rates for high-income earners (45% income tax; 1.5% Medicare Levy; 1% Medicare Surcharge). (11) As a result of these factors, differences in expenditure are almost doubled in terms of the salary required to service that expenditure. For example, consider the case of the purchaser couple without children, when one adult is earning $116,000 and the other has no income (see Table 2). In Brisbane, this equates to an expenditure of approximately $73,500 per year, with an equivalent expenditure in Sydney of approximately $98,000 per year; an increase of about $24,500. The Sydney salary required to achieve this expenditure is found to be about $165,000, or $39,000 more than the Brisbane salary, representing a 42 per cent increase. These considerable variations are mainly evident for high-income households and only in Sydney.

Percentages differences provide a clearer picture of the differences between capital cities and how they vary. Summarising the findings, we examine the differences in equivalent salaries first according to household type, then to income level and finally to dwelling type.

Table 3 provides the range and mean of relativities between Brisbane and other capital cities for different household types for September quarter 2006. The Table reinforces the previous observation about the higher costs in Sydney. The bottom row of the Table shows that compared with Brisbane an average increased private income of 24 percent is required for households living in Sydney, an average increase of 7 percent in Melbourne, an average increase of 4 percent in Perth, and incomes in Adelaide are required to be on average 7 percent less than Brisbane. Table 3 also indicates that at times salaries in Adelaide need to be close to half that of Brisbane salaries to obtain the same living standard. This occurs when the Adelaide households can supplement private income with social security benefits, as in the case in the shaded cell in Table 2. Looking down the columns of Table 3, it appears that the larger the household size, the larger the variation between Brisbane and another capital city. This could be partly related to the greater average income for a larger family--given government benefits--than for smaller families.

Table 4 reports both the range and the mean variation in equivalent salaries, relative to Brisbane, for households of different private income levels for September quarter 2006. For example, the Table indicates that households with a total private household income of $58,000 in Brisbane require between a 1 percent and a 42 percent increase in salaries in Sydney (with an average increase of 21 percent) to achieve the same standard of living. The variation results from varying the household type. In contrast, an average increase of 8 percent is required at this income level in Melbourne and a 2 percent and a 7 percent decline is, on average, required in Perth and Adelaide respectively to maintain the same standard of living.

Table 4 also suggests that as total household income rises, the increased equivalent household income in Sydney is relatively stable, whereas the disparity between Brisbane and Adelaide and Melbourne declines. In contrast, disparity between Brisbane and Perth households appears to increase with household income.

Table 5 provides the breakdown in relativities according to dwelling type. It shows that the dwelling type and when a house was purchased can make a significant difference to the relativities between cities. For example, in September quarter 2006, privately renting in Sydney requires an average increased salary of 15 percent relative to Brisbane, which is identical to the 15 percent increased salary required for those who purchased a dwelling in March quarter 1993, whereas purchasing in September quarter 2006 requires a massive 41 percent increase in salary to maintain the same standard of living. As this example illustrates, the greatest disparities between Brisbane and other capital cities are for recent home purchasers. (The Adelaide rental market is the exception.) However, care should be taken when interpreting this fact. This outcome is not because recent mortgages are larger than private rents, but reflects the growth in the disparities between cities in mortgage/purchase prices.

This brings the discussion to the final point of analysis. How have these relativities changed over time? The discussion thus far has considered the analysis for the September quarter 2006. How do the findings at this point in time compare with earlier time periods? As indicated earlier, this exercise was also undertaken for two earlier periods: September quarter 2005 and March quarter 2003. Figure 1 provides a graphical summary of the change in average relativities between Brisbane and other capital cities across all household types and living standards in early 2003, late 2005 and late 2006.

It can be seen from Figure 1 that the average differential between Brisbane and Sydney remained reasonably constant at about 40 percent from early 2003 to late 2005, but in the year from late 2005 to late 2006 that differential significantly dropped to 27 percent. (12) In the same three and half year period, the relativities between Brisbane and Melbourne have also declined somewhat--from 15 to 7 percent. However the narrowing of the differential occurred in the period 2003 to 2005, instead of the later period for Sydney. In contrast, the equivalent salary differential between Brisbane and Adelaide has remained remarkably constant at 7 percent less than Brisbane. Moving in the opposite direction is the differential between Brisbane and Perth. From 2003, when Perth was similar to Adelaide, Perth became more like Brisbane, and in 2006 overtook Brisbane in incomes required to achieve the same standard of living.

[FIGURE 1 OMITTED]

Clearly shifts in housing costs are a major explanation for the relativities and changes in relativities. Indeed, since September 2006 and December 2007, median house prices in Sydney relative to Brisbane have dramatically declined from 58% to 34%. (13) However, childcare costs can also contribute to differential costs. Calculations undertaken by the author suggest that Sydney childcare costs are amongst the cheapest of the cities considered in this report, whereas Melbourne is the most expensive. (14) Food costs are also a significant component of household budgets, and this most expensive in Sydney.

Discussion

For commentators and scholars watching real estate movements, the differentials between cities would not come as a surprise. No doubt, the differences in housing costs--and thus the salaries required to attain a similar standard of living--result from a number of regional factors, including economic growth levels, population growth and the supply of land. The question that I wish to briefly consider is what should be the public policy response to these differentials, if any.

As noted at the beginning of this paper, Australia's public policy setting largely does not differentiate on the basis of location. Clearly we have differences between States in their policies and differences occur in local governments by charging rates on the basis of land value. The benefits and taxation systems include a Remote Area Allowance and a Zone Tax Offset, but no distinctions in costs are made between cities. Research showing the differential costs of households living in different parts of Australia raises the question as to whether to introduce more targeted policies to enhance equity, particularly for lower income households.

If it costs more to live in Sydney and less in Adelaide, then it would seem to follow that policy could be smarter in addressing these differentials. For example, instead of the current nation-wide Rent Assistance policy, a regionally-based Rent Assistance for low income families could be introduced. This could also be seen to encourage and enable the unemployed to move towards areas of high job growth, and away from low-cost, low-employment areas and would be consistent with the Federal government's welfare to work agenda. Another response would be to introduce tax allowances for high cost areas, or for employers to offer loadings for living in the more expensive areas.

Such policy responses are not without difficulties. Increasing incomes (by employer loadings or government benefits) may only have the effect of increasing the differential. If people were encouraged to go to high cost areas, this would generate more demand, and potentially increase housing, child care and other localised costs. Indeed, it is worth noting that Canberra currently has one of the highest housing and childcare costs, arguably driven up by one of the highest average salaries in a supply-restricted market. The flip side is that encouraging people to move away from low-cost areas may actually (further) undermine the local economy of that area. The case of Tasmania is a case in point. Rather, if areas of low cost reflect lower economic outcomes, then higher relative wages may be a strategy to encourage population growth and hopefully economic growth.

Another problem with regionally-targeted benefits to counter geographical variation in prices is one of perception. Whilst such policies are aimed at enhancing equity, the public may well view giving benefits or tax breaks to people in high-cost areas as unfair, because policy treats people differently (cf Henman 2005).

A final problem with a public policy response to geographical price differences is that the differential identified between cities may reflect locational benefits not measured in research. For example, compared to Brisbane, residents in Sydney are closer to beaches, better restaurants and other niche services, and arguably may have better public transport. These factors are not included in the above comparisons, and may explain why people are prepared to pay more for housing, child care and so on in Sydney. Of course, there are also other unmeasured benefits in Brisbane, over Sydney, such as reduced pollution and traffic congestion, lovely weather and a relaxed lifestyle.

The above comments clearly demonstrate that any public policy responses to differential costs are problematic. That is not to say that nothing can or should be done, but that policy needs careful political, technical and social consideration.

References

Australian Bureau of Statistics (ABS), Average Retail Prices of Selected Items, Eight Capital Cities, Catalogue No. 6403.0, Canberra.

Australian Bureau of Statistics (ABS), Labour Price Index, Australia, Cat. No. 6345.0, Canberra.

Australian Bureau of Statistics (ABS), Consumer Price Index Standard Data Report: Capital City Index Numbers (a) by Expenditure Class, Catalogue No. 6455.0.55.001, ABS, Canberra.

Australian Bureau of Statistics (ABS) (2007), Household Expenditure Survey 2003-04: Broad Expenditure Groups by Quintile by Selected Age Ranges and Selected Household Family Composition, unpublished, specially produced dataset data, ABS, Canberra.

Bray, J.R. (2002), Social indicators for regional Australia, Policy Research Paper No. 8.

Castles, I. (1997), Review of the OECD-Eurostat PPP Program, STD/PPP (97) 5, Paris: OECD, http://wwwl.oecd.org/std/ecastle.pdf

Centrelink (2007), Remote Area Allowance, Canberra, http://www.centrelink. gov.au/internet/internet.nsf/payments/remote_area.htm

Department of Family and Community Services (FaCS) (2005), 2004 Census of Child Care Services, FaCS, Canberra, http://www.facsia.gov.au/internet/ facsinternet.nsf/childcare/04_census.htm

Department of Social Security (DSS) (1996) Overview of the Australian Private Rental Market, Policy Research Paper No. 72, Canberra: DSS.

Gregory, B. and Hunter, B. (1996), 'Increasing Regional Inequality and the Decline of Manufacturing,' in P. Sheehan et al. (eds), Dialogues on Australia's Future, Centre for Strategic Economic Studies, Melbourne.

Henman, P (2001), Updating Australian budget standards costs of children estimates, Policy Research Paper No.7, Department of Family and Community Services (FaCS), Canberra.

Henman, P. (2005), 'E-government, targeting and data profiling: policy and ethical issues of differential treatment', Journal of E-government, 2(1): 79-98.

Henman, P. (2007a), 'Updated costs of children using Australian budget standards' in Costs of Children: research commissioned by the Ministerial Taskforce on Child Support, Occasional Paper No. 18, Department of Families, Community Services and Indigenous Affairs, Canberra, pp. 1-54.

Henman, P. (2007b), Equivalent salaries in Australian capital cities: an update, Brisbane: Queensland Department of Employment and Industrial Relations, http://www.dir.qld.gov.au/pdf/ir/qdirreport2007.pdf.

Henman, P. and Mitchell, K. (2001), 'Estimating the Costs of Contact for Non-resident Parents: a budget standards approach', Journal of Social Policy, 30(3), 495-520.

Lawless, P. (2006), 'Area-based urban interventions: Rationale and outcomes: The New Deal for Communities program in England', Urban Studies, 43(11), 1991-2011.

OECD (2002) Purchasing power parities--measurement and uses, Statistics Brief No. 3, Paris: OECD.

Real Estate Institute of Australia (REIA) (2007) Real Estate Market Facts, March Quarter 2007, Canberra: REIA.

Saunders, P. (1998) Using budget standards to assess the well-being of families, SPRC Discussion Paper No. 93, Sydney: University of New South Wales.

Saunders, P., Chalmers, J., McHugh, M., Murray, C., Bittman, M. and Bradbury, B. (1998), Development of Indicative Budget Standards for Australia, Policy Research Paper No. 74, Department of Social Security, Canberra.

Saunders, P., Patulny, R. and Lee, A. (2004), Updating and Extending Indicative Budget Standards for Older Australians, SPRC Report 2/04, Social Policy Research Centre, University of New South Wales, Sydney.

Siminski, P. and Saunders, P. (2004) Accounting for housing costs in regional income comparisons, SPRC Discussion Paper No. 135, Sydney: University of New South Wales.

Swan, W. (2005), Postcode: the splintering of a nation, Pluto Press, North Melbourne.

Vinson, T. (2004), Community Adversity and Resilience, Jesuit Social Services, http://acl.arts.usyd.edu.au/jss/.

(1) Examples of Australian geographical socio-economic research abound. For example, see Gregory and Hunter (1996) as an important early contribution, Bray (2001) in the government literature, Vinson (2004) as an important recent study, Swan (2005) in the popular political literature, and almost any article in the journal People and Place.

(2) Saunders {1998) provided an earlier and much limited derivation the capital city budget standards.

(3) Another constraint is information about relative quality. See the later discussion on housing.

(4) See Saunders et al (1998, 628-631) for a discussion of these issues and Saunders et al (2004) for the creation of a higher "comfortably affluent but sustainable' living standard for retirees.

(5) In this respect, it is notable that several public policy changes, especially the halving of capital gains tax, has contributed to the growth in housing demand in the period from 2000.

(6) Saunders et al (1998) cautioned against updating the Australian budget standards beyond 5 to 7 years solely on prices and without a review of the content of the basket of goods and services. Ideally, such an update should be done. However, while updating a basket's contents would change the cost of a household budget standard, the relativities between cities would not significantly change.

(7) The specially constructed data set from the HES 2003-04 involved three different household types each with reference adult of age less than 65: single adult households; couple only households; and couple with dependent children households. Each of these three populations were segmented into quintiles according to household income (not expenditure). The average expenditure for different expenditure categories was calculated for the households in each quintile. Budget standards for high income sole parents were based on the couple with children relativities.

(8) Straight linear regression and quadratics were also considered, but it was found that the log-linear equation produced the most accurate approximation to the expenditure points, particularly at lower expenditures where small errors have greater significance.

(9) The tax-benefits model was designed for single and couple adult-headed households, with varied number and age of children. An input of the private income of each adult in the household is used to calculate the impact of relevant taxation and social security policy for each of the three time periods examined and the consequent household disposable income. Policies incorporated into the model were: income taxation on wages/salaries; Medicare Levy and Medicare Surcharge; Low Income Rebate; Dependent Spouse Rebate; Family Tax Benefit Parts A and B; Child Care Benefit; Child Care Rebate (for the September quarter 2006 only); and Income support payments for households with low income, namely, Parenting Payment (Partnered); Parenting Payment (Single); Newstart Allowance; Rent Assistance; and Allowance Rebate.

(10) All other calculations can be obtained from Henman (2007b).

(11) For families with children, this marginal tax rate can increase by a further 30 per cent in the tapering out of Family Tax Benefit Part A around the vicinity of private household incomes of $90,000 to $110,000.

(12) Data for 2006 in Figure 1 are slightly different to those presented in Table 3. This is because the analysis for earlier years did not contain consideration of a sole parent household. Figure 1 maintains the same population to ensure comparisons are like with like.

(13) Calculated from data from REIA's Market Facts.

(14) This finding is based on updating government administrative data on the gross childcare costs from late 1998 (Henman (2001: 17) to the present using ABS's CPI childcare index for each capital city. This finding does seem at odds to the government's Child Care Census which found that the ACT and NSW had the highest costs (FaCS 2005).

Table 1: Budget Standards Household Types

 Low Cost Low Cost+ MBA

 Private renters Purchasers Private renters

Single adult 2 bed unit; FT 2 bed unit; FT 2 bed unit; FT

Couple no 2 bed unit; FT/ 2 bed unit; FT/ 2 bed unit;
 children NILF NILF a) FT/NILF
 b) FT/FT

Couple with one 3 bed unit; FT/ 3 bed unit; FT/ 3 bed house;
 child NILF NILF a) FT/NILF
 b) FT/FT

Couple with two 3 bed house; 3 bed house; 4 bed house;
 children FT/NILF FT/NILF a) FT/NILF
 b) FT/FT

Sole parent with 3 bed house; 3 bed house; 4 bed house;
 two children FT FT FT

 MBA+ Luxury

 Purchasers Purchasers

Single adult 2 bed unit; FT 2 bed unit; FT

Couple no 2 bed house; 2 bed house;
 children a) FT/NILF a) FT/NILF
 b) FT/FT b) FT/FT

Couple with one 3 bed house; 3 bed house;
 child a) FT/NILF a) FT/NILF
 b) FT/FT b) FT/FT

Couple with two 4 bed house; 4 bed house;
 children a) FT/NILF a) FT/NILF
 b) FT/FT b) FT/FT

Sole parent with 4 bed house; 4 bed house;
 two children FT FT

Key: FT-employed full-time; NILF=not in the labour force;
MBA=modest but adequate living standard.

Table 2: Equivalent Salaries and Disposable Incomes--Couple only
households ($'000s/year, rounded to nearest $100)

Brisbane Sydney

Salary1 Salary 2 Disp Inc.* Disp Inc.* Salary 1 Salary 2

Couple--One full-time--Private renter

$40,600 $0 $32,000 $33,900 $43,600 $0
$58,000 $0 $43,100 $45,600 $61,900 $0
$87,000 $0 $60,300 $64,100 $94,200 $0
$116,000 $0 $73,500 $78,500 $125,800 $0

Couple--One full-time Purchaser 2006

$40,600 $0 $32,000 $42,900 $57,700 $0
$58,000 $0 $43,100 $56,400 $79,600 $0
$87,000 $0 $60,300 $79,400 $127,400 $0
$116,000 $0 $73,500 $98,000 $165,000 $0

Couple--Both full-time--Private renter

$40,600 $40,600 $60,900 $64,800 $43,700 $43,700
$58,000 $29,000 $65,000 $69,300 $62,700 $31,300
$58,000 $58,000 $80,600 $86,500 $62,700 $62,700
$87,000 $58,000 $97,300 $104,800 $95,200 $63,400
$116,000 $58000 $112,200 $121,400 $127,000 $63,500

Couple--Both full-time--Purchaser 2006

$40,600 $40,600 $60,900 $80,100 $57,500 $57,500
$58,000 $29,000 $65,000 $85,800 $83,800 $41,900
$58,000 $58,000 $80,600 $108,200 $181,500 $81,500
$87,000 $58,000 $97,300 $132,700 $126,300 $84,200
$116,000 $58,000 $112,200 $155,200 $170,600 $85,300

Melbourne (^) Adelaide

Disp Inc.* Salary 1 Salary 2 Disp Inc.* Salary 1 Salary 2

Couple--One full-time--Private renter

$31,500 $39,800 $0 $29,600 $22,000# $0
$43,500 $58,700 $0 $40,000 $53,200 $0
$62,200 $90,500 $0 $55,700 $78,400 $0
$76,400 $121,700 $0 $67,500 $104,300 $0

Couple--One full-time Purchaser 2006

$32,600 $41,500 $0 $31,300 $34,100 $0
$44,600 $60,300 $0 $41,400 $55,300 $0
$63,200 $92,400 $0 $57,700 $82,100 $0
$77,500 $123,800 $0 $70,500 $110,200 $0

Couple--Both full-time--Private renter

$62,200 $41,600 $41,600 $56,300 $36,800 $36,800
$66,600 $59,800 $29,900 $60,000 $52,700 $26,300
$83,600 $60,400 $60,400 $73,900 $52,500 $52,500
$101,600 $91,700 $61,100 $88,500 $77,600 $51,700
$117,800 $122,600 $61,300 $101,600 $102,900 $51,500

Couple--Both full-time--Purchaser 2006

$63,300 $42,500 $42,500 $58,300 $38,500 $38,500
$67,700 $60,900 $30,500 $62,300 $55,100 $27,600
$84,700 $61,300 $61,300 $77,600 $55,500 $55,500
$102,700 $92,800 $61,900 $94,000 $83,500 $55,600
$118,900 $123,900 $62,000 $108,900 $111,800 $55,900

Perth

Disp Inc.* Salary 1 Salary 2

Couple--One full-time--Private renter

$31,400 $34,600 $0
$42,300 $56,700 $0
$61,400 $89,000 $0
$77,000 $122,800 $0

Couple--One full-time Purchaser 2006

$34,600 $44,600 $0
$49,300 $67,700 SO
$72,100 $113,300 $0
$89,600 $147,300 $0

Couple--Both full-time--Private renter

$62,100 $41,600 $41,600
$66,900 $60,100 $30,100
$85,800 $62,200 $62,200
$106,600 $97,000 $64,700
$125,600 $132,100 $66,100

Couple--Both full-time--Purchaser 2006

$71,200 $48,800 $48,800
$76,700 $73,000 $36,500
$97,500 $71,700 $71,700
$119,600 $111,100 $74,100
$139,400 $149,000 $74,500

* Disposable income is after tax and a 5% superannuation contribution
from gross private income

(^) Linear regression without a logarithm was used for Adelaide and
Perth due to distortions at the higher income households with the
logarithm regression.

Shaded cells indicate that private income is supplemented by Newstart
Allowance and, where relevant, Rent Assistance.

Note: Private income is supplemented by Newstart Allowance and, where
relevant, Rent Assistance indicated with #.

Table 3: Range and Average Percentage salary change to obtain
equivalent salary relative to Brisbane, by household type,
September quarter 2006

Household type Sydney Melbourne

Single adult Range Average 7% to 43% 17% -8% to 10% -0%
Couple only Range Average 7% to 47% 26% -2% to 7% 4%
Couple + 1 Range Average 9% to 52% 28% +0% to 14% 4%
Couple + 2 Range Average 1% to 56% 32% -6% to 32% 8%
Sole parent + 2 Range Average 13% to 56% 37% 3% to 35% 22%

Overall Range Average 1% to 56% 24% -8% to 35% 7%

Household type Adelaide Perth

Single adult Range Average -27% to -4% -17% -16% to 29% -0%
Couple only Range Average -46% to +1% -10% -15% to 30% 13%
Couple + 1 Range Average -15% to 17% -5% -13% to 37% 7%
Couple + 2 Range Average -46% to +1% -6% -18% to 32% 7%
Sole parent + 2 Range Average -20% to -5% -10% -37% to 66% +1%

Overall Range Average -46% to 17% -7% -37% to 66% 4%

Table 4: Range and average salary change to maintain living standard
equivalence, by total private household income, September quarter
2006

Household private income Sydney Melbourne
 in Brisbane

$40,600 Range 7% to 56% -2% to 33%
 Average 28% 10%
$58,000 Range 1% to 42% -3% to 35%
 Average 21% 8%
$87,000 Range 8% to 56% -6% to 31%
 Average 28% 9%
$116,000 Range 7% to 42% -8% to 13%
 Average 20% 4%
$145,000 Range 9% to 45% -1% to 11%
 Average 22% 4%
$174,000 Range 8% to 47% -6% to 12%
 Average 22% 3%

Household private income Adelaide Perth
in Brisbane

$40,600 Range -46% to 17% -37% to 66%
 Average -10% 1%
$58,000 Range -17% to +1% -31% to 28%
 Average -7% -2%
$87,000 Range -23% to 1% -27% to 30%
 Average 6% 5%
$116,000 Range -27% to -1% -26% to 27%
 Average -8% 3%
$145,000 Range -11% to -1% -5% to 31%
 Average -6% 9%
$174,000 Range -11% to -2% -6% to 37%
 Average -8% 10%

Table 5: Range and average salary change to maintain living standard
equivalence relative to Brisbane, by dwelling type, September quarter
2006

Dwelling type Sydney Melbourne

Private renter Range 1% to 34% 8% to 35%
 Average 15% 5%
Purchaser 2006 Range 15% to 56% -3% to 32%
 Average 41% 8%
Purchaser 1993 Range 8% to 31% -6% to 26%
 Average 15% 7%

Dwelling type Adelaide Perth Overall

Private renter Range -46% to +1% -37% to 14% 1%
 Average -10% -6%
Purchaser 2006 Range -46% to 17% -2% to 66% 16%
 Average -7% 20%
Purchaser 1993 Range -20% to +1% -27% to 14% 4%
 Average -4% -3%