The methodology of budget standards

A budget standard is based on three forms of information: normative data (and judgements) provided by experts that describe human needs and the commodities (or activities) required to satisfy these needs at a minimal standard, behavioural data that describe what people actually spend their money on and experiential data that capture the details of how low-income individuals actually spend their money. Data relevant to these latter two elements are derived from social surveys and focus group discussions, respectively, since these are the appropriate analytical tools for accessing the required information. All three elements are important if the budgets are to reflect existing knowledge about human needs, expenditure patterns and shopping decisions and practices. There is, however, on-going debate about how much weight should be assigned to the different forms of evidence, with some studies (particularly in the UK) giving a greater deliberative role to the focus group discussions and placing more weight on those findings.^{Footnote 1} However, most would agree with Reference DeemingDeeming (2017) who has argued that ‘… methodological pluralism should prevail. There is strength in the different budget approaches; all are varied, and arguably, needed’ (p. 44).

Australian experience with using budget standards in the wage-setting context has not resulted in agreement on how best to apply the methodology, although there is broad agreement that the approach has much to contribute. The Australian Catholic Council for Employment Relations (ACCER) has argued, for example, that

The Australian Fair Work Commission (FWC, 2014) has also acknowledged the value of budget standards, albeit more guardedly, noting,

Nevertheless, support for the approach in other contexts has waned in the face of criticism that it relies heavily on arbitrary judgements that exert a powerful influence on the results produced. However, it is difficult to see how any attempt to derive benchmarks that can be used for assessing adequacy can fail to contain normative elements. This view is reinforced by studies conducted by Australian Government agencies over the past three decades that have consistently defined adequacy in terms of ‘providing a basic acceptable standard of living, accounting for prevailing community standards’ (Reference HarmerHarmer’s (2009) Pension Review: xii–xiii; see also Reference Saunders and WongSaunders and Wong, 2011, for a wider review). Giving practical effect to this conceptual definition involves giving operational meaning to phrases such as ‘basic acceptable’ and ‘prevailing community standards’ that are inherently normative and thus require judgements to be made.

The fact that any such judgements will always be open to debate does not negate the need for budget standards research. Its value lies in its unique ability to highlight the choices that have to be made, set out possible alternatives, provide a systematic framework for choosing among them and examining the impact of different choices on the results produced. These features are inherent in other approaches to setting adequacy benchmarks such as poverty lines, although their importance is often ignored or obscured by the use of simple judgements that drive the results produced. Thus, for example, shifting the poverty line from 50% to 60% of median income will result in a decline in the adequacy of social benefits, but only because the judgement used to identify poverty has changed.

The budget standards approach has been modified to address these and other weaknesses, to incorporate improved research methods and better data improvements, and to reflect the experience gained with their development and application. The application described here continued to use the three basic ingredients identified above – normative, behavioural and experiential. It did not, however, assign a greater role to focus groups but instead utilised the better range and quality of behavioural data that exist today compared with when the original study was conducted two decades ago. Further details of what this involved are provided in the following discussion.

Developing the new budget standards: Similarity and difference

Users of budget standards understandably want the estimates to be relevant to current circumstances, that is, to reflect ‘prevailing community standards’. Updating the original SPRC budgets to take account of changes in consumer prices will ignore improvements in overall living standards or changes in item availability, retail and shopping practices and other behavioural norms.^{Footnote 2} One commonly used measure of overall living standards – household disposable income per head of population (HDIPC) – increased almost two and a half times (or by 248%) between 1994–1995 (when the original SPRC budgets were priced) and 2013 (when those presented here were initially priced). It is hard to believe that a change of this magnitude in the average level of economic prosperity will not have had a profound impact on what is regarded as an acceptable minimum standard of living and thus on the items required to support that standard.

Adjusting budgets in line with movements in the Consumer Price Index (CPI) and its components will not take account of these changes (except indirectly, as the CPI weights are revised) with the result that the updated budgets will slowly drift apart from current realities. This divergence is particularly an issue in relation to the setting of the minimum wage, where contemporary relevance is critical. It was recognised by the FWC (2014), which followed its positive assessment of the role of budget standards research cited earlier by noting that

This assessment highlights the pressing need to review and revise (not just update) the original SPRC budget standards.^{Footnote 3}

The initial aim of the new research was to limit the extent of any major divergence from the original study in order to maintain comparability. In some areas, this was straightforward, including, for example, the use of ‘rules of thumb’ (see Reference Saunders, Chalmers and McHughSaunders et al., 1998: 67–73) to provide a consistent approach when deciding whether or not to include an item. In other areas, replicating what was done before was more problematic – often because a review of the methods used in the original study suggested that a new approach was needed or because the availability of better data allowed improvements to be introduced.

The most significant change relates to the standard to which the budgets were tied since this determines which items to include and how much they cost. Two separate standards were used in the earlier SPRC study, both drawing on the work of American economist Reference WattsHarold Watts (1980). The first, the Modest but Adequate (MBA) standard, was intended to provide access to available economic and social opportunities, corresponding to a level of decency between survival and affluence and falling around the median (mid-point) of living standards in the community as a whole. The second, the Low Cost (LC) standard, was intended to capture a minimal standard of living that requires the frugal use of resources but still allows a degree of economic and social participation, in practice, falling at around one-half of the median. Translating these two standards into actual budgets proved difficult because of the over-determination implicit in specifying both what each standard represents and where it falls in the overall distribution.^{Footnote 4}

The MBA and LC standards were thus replaced by the Minimum Income for Healthy Living (MIHL) standard that has emerged from the UK public health literature (Reference Morris and DeemingMorris and Deeming, 2004; Reference Morris, Deeming and WilkinsonMorris et al., 2010 see also Reference O’Sullivan and AshtonO’Sullivan and Ashton, 2012, for an application to New Zealand). The basic idea of healthy living is that all family members (adults and children) are able to maintain the healthy lifestyle required to participate fully in their roles as workers, consumers and citizens. The MIHL standard thus has relevance to the social inclusion agenda because poor health and disability are barriers to economic and social participation while also being consistent with broader individual and societal objectives relating to the promotion and maintenance of good health. This approach is consistent with that adopted in the European Union Reference Budgets study, where the targeted living standard is defined as having

Clearly, this does not mean that the MIHL standard is the only option available, although the limitations of the MBA and LC standards are now generally accepted among budget standards researchers and the MIHL standard is one of the few alternatives that has been used. The estimates that follow thus represent an initial attempt to examine the strengths and pitfalls associated with applying the MIHL standard in the Australian context.

Once the standard has been set, the budgets themselves have to be constructed by the painstaking task of identifying and costing items, many of them at the individual level but others at the household level, and then aggregating to obtain the family budget.

Another important difference in approach involved developing budgets for ‘average’ families rather than (as before) ‘hypothetical’ families. The use of the latter concept in the original study necessitated specifying where the family lived and, from this, identifying the type and size of dwelling (and its cost) and constructing a transport budget by mapping the various trips it would have to make (to local schools, health facilities, employment offices and so on). Both aspects created practical problems and made the final budgets too dependent on the specifics of the chosen area, such as the nature of the local housing market and the availability of public transport. These problems made it difficult to refute claims that the budgets were not able to serve as general benchmarks, even for families of a given type. The use of the average family avoids these limitations, and although it gives more weight to the behavioural element of the budget (and to data derived from social surveys), this was seen as a reasonable price to pay to achieve greater relevance and applicability.

A concerted effort was made to engage directly with low-income people through focus groups run with unemployed and low-paid individuals. The aim was to better ground the budgets in the realities of everyday experience for those at the cutting edge of social disadvantage by getting feedback on the budgets (both the items included and their cost) and ensure that they reflected what is happening ‘on the ground’. While this approach proved to be very helpful in relation to focus groups with unemployed people, it was far harder to recruit participants from among low-paid workers because many were working unstable and unpredictable hours and could not commit to a specific meeting time, while finding a common meeting place presented further challenges. In contrast, the unemployed were required to meet regularly with their job service provider and (with considerable assistance from these providers) were able to be brought together on two separate occasions for the relevant discussions. However, although the focus group input ended up diverging from what was planned, the discussions that did take place provided a range of important information that was used to develop the new budgets (as explained further below).

The variation in housing costs referred to above is a major obstacle to developing a representative housing budget. While variation across dwelling types is to be expected in a budget standards context (because larger families require dwellings with more rooms), the cost of a given dwelling also varies with tenure status and, importantly, location. The former can be avoided by focusing on one tenure type (in this case renting privately, where market data are plentiful and reliable), while the latter creates problems for all adequacy benchmarks. In the earlier study, families were located in Sydney and this led to a chorus of criticism that the budgets were exaggerated by Sydney’s high housing costs. In order to avoid this problem, this time the budgets were first derived excluding housing and were then ‘grossed-up’ to include housing costs based on the average rents for different dwellings located in different suburbs in Australia’s three largest cities – Sydney, Melbourne and Brisbane. The rent estimates were based on data extracted from the market facts reports produced by the Real Estate Institute of Australia (REIA, 2016).^{Footnote 5} Having first established the dwelling type for each family based on the number of bedrooms needed, the researchers then assigned families to a suburb within each city and added average rents in those suburbs to the non-housing budgets. This approach is transparent and leaves it open to others to use different assumptions when applying the REIA data without having to discard the effort that has gone into constructing the other budget components.^{Footnote 6}

Two offsetting factors impacted the choice of lifetimes assigned to long-lasting (durable) items in the budgets. The first is built-in obsolescence, which over time reduces lifetimes and thus increases the weekly cost of durables. Against this, the focus groups provided many examples of strategies used to extend the lifetimes of items, particularly items of children’s clothing, where free exchange was a common way of handing on (and receiving) clothing that children had outgrown. Lifetimes for household white goods and furniture were also assumed to vary with the presence of children, with items assumed to wear out more quickly when children are present. The balance between these competing forces was that, overall, the lifetimes included in the new study are shorter than those used originally.

The new budgets are minimal and conservative, being based on the lowest cost option whenever alternatives were available. Experience with the use of the original budgets dictated that ‘tight’ budgets would gain greater credibility than those that erred on the side of generosity – as long as this did not prevent the family and its members from achieving the MIHL standard.

The new budget standards

New budget standards were derived at the MIHL standard for families where the main breadwinner was either in full-time work (in part-time work in the case of the sole parent) and being paid the minimum wage, or out of work and receiving NSA.^{Footnote 7} For convenience, these two groups are referred to henceforth as the low-paid and unemployed families (or budgets). Budgets were derived for five basic family types: single people (males and females separately, although only the average, gender-neutral version is discussed here); couples without children; couples with one and two children; and a (female) sole parent with one child.^{Footnote 8} Male adults are assumed to be aged 40, female adults aged 35, while the first child in each case is a 6-year-old girl and the second child a 10-year-old boy.^{Footnote 9} Families with either no or only one adult in (low-paid) employment are located at the bottom of the working-age income distribution although it is difficult to provide exact estimates because of the small numbers involved in, for example, the 2009–2010 Household Expenditure Survey (HES), that have similar demographic characteristics to those assumed here (see Reference Saunders and BedfordSaunders and Bedford, 2017: Table C.1)

The new budgets were originally priced in the latter half of 2013 in leading national supermarkets such as Woolworths and Kmart, although the budgets were refined as items were adjusted, added or removed between then and early 2016. Each time this was done, new prices were obtained, but they were then deflated back to the original pricing window. Once the entire budgets were finalised, they were updated to the June Quarter of 2016 in line with movements in the relevant CPI components.

The new low-paid and unemployed budgets and budget shares (‘grossed-up’ to include housing costs) are presented in Tables 1 to 3 for each family type. The budgets highlight the importance of housing costs, which account for between 39% and 53% of the low-paid budgets and between 36% and 52% of the unemployed budgets.^{Footnote 10} In all cases, the budget shares devoted to housing exceed the 30% rule that is commonly used to identify housing stress among low-income families (see Reference Rowley and OngRowley and Ong, 2012; Reference YatesYates, 2007). The other areas that contribute significantly to the total budgets are food, household goods and services (which includes electricity and other utility bills) and transport.^{Footnote 11} Education and health contribute only a modest amount to the family budgets, in both instances because most of the costs are borne by the public purse not the family wallet.^{Footnote 12}

Table 1. New budget standards for a single person and a couple with no children, June 2016.

Table 2. New budget standards for couples with one and two children, June 2016.

Table 3. Budget standard for a sole parent with one child (6-year-old girl) (AUD/week).

In all cases except the single adult, the budgets for those who are unemployed and on NSA are around 20% lower than those where the family breadwinner is in full-time work and receiving the minimum wage. For single people, the relativity is around 27% and is largely a reflection of the greater difference in housing costs facing single people. Aside from this, the fact that the budget relativities are similar across family types suggests that the underlying standards are broadly constant for each of the two labour force states. These are the relativities required to ensure that families at the two standards (low-paid and unemployed) are able to achieve the same level of adequacy. Other considerations (e.g. the need to maintain appropriate work incentives for those in and out of work) might dictate a different relativity, but such decisions cannot be made without access to the evidence about what the alternatives imply for the structure of both adequacy and incentives (and the costs involved).

Comparisons with existing benchmarks and safety net provisions

Table 4 compares the grossed-up budget standards with the relative poverty line that is most commonly used in Australian research studies and/or policy discussions, set at one-half of median (equivalised) income (see Reference Cassells, Dockery and DuncanCassells et al., 2014; Reference Saunders, Wong and BradburySaunders et al., 2016; Reference WilkinsWilkins, 2008). The new budget standards estimates for low-paid couple families are between 22% and 27% above the median income poverty line, although these gaps are larger for single adults (35.4%) and the sole parent with one child (46.9%). The fact that the new budget standards are well above the poverty line for those in low-paid work is not surprising, since the primary goal of Australia’s wage-setting system since its inception has been to protect workers against poverty, and this involves ensuring that the minimum wage is above the poverty line.

Table 4. Comparing the grossed-up budget standards with existing poverty lines, June Quarter 2016 (AUD/week).

Table 4 also indicates that the grossed-up budgets for unemployed single people and couple families are all very close to the corresponding median income poverty lines. This finding confirms that the standard implicit in the new unemployed budgets is very conservative, corresponding to how much is needed to attain the absolute minimum levels of living associated with poverty. The main divergence from this pattern occurs for the sole parent family, where the new budget standard is close to 20% above the median income poverty line, that is, at a level closer to 60% of median income than 50%. This is despite that fact that the age of the child is such that child care costs are kept to a minimum, although the sole parent family (unlike the other unemployed families) was assigned a car to allow the parent to better juggle her many responsibilities.

Overall, the comparisons in Table 4 indicate that if a poverty line tied to median income is used as the benchmark, the new budget standards would be more than adequate to support a poverty line standard of living if the family has at least one member in low-paid employment, even if only on a part-time basis, as is the case of the sole parent. However, this is not the case for most unemployed families, where the level of NSA does not provide effective protection against poverty.

Table 5 compares the new budget standards with the incomes generated by the Australian social safety net and provides a series of evidence-based benchmarks for assessing the adequacy of these provisions. For each family type, the existing rules and eligibility criteria have been applied to calculate what level of income would be received as cash in hand (or disposable income) if the family (male) head is either working full-time and receiving the minimum wage, or is unemployed, looking for work and in receipt of NSA.^{Footnote 13}

Table 5. Comparisons of the grossed-up budget standards with existing safety net incomes (AUD per week, June Quarter 2016).

Source: Tables 1 to 3 and see main text.

The safety net incomes shown in Table 5 reflect the incomes received by low-paid workers under the April 2016 minimum wage award (an hourly rate of AUD17.29 or AUD656.90 a week) or (for the unemployed) the prevailing maximum base rate of NSA (equal to AUD506.85 per fortnight or AUD253.40 per week) or (for the sole parent) the base rate of Parenting Payment, Single. The calculations also assume that families receive any additional benefits for which they are eligible, including (where appropriate) Rent Assistance, the Energy Supplement, Income Support Bonus, Family Tax Benefits Parts A and B and the associated Supplement and Energy Supplements and (for the sole parent family) Telephone Allowance. Income tests have been applied where this affects the payments made to families in paid work or receiving multiple payments, and all incomes are expressed net of any income tax liabilities, after factoring in where relevant the Low Income Tax Offset and the Senior Australians and Pensioners Tax Offset (for the sole parent family) and the Medicare levy.

These comparisons indicate that the safety net is providing an adequate income floor for low-paid individuals and sole parents in part-time work but not for couples without children or with one child, where the shortfall varies between AUD39 and AUD89 a week. Even in those two cases where the gaps are positive, they are not large enough to cover the higher rents that would have to be paid by those living in Sydney compared with the three-city average rents incorporated into the budget standards.^{Footnote 14} In light of the fact that the budget standards are deliberately conservative, the fact that existing safety net provisions only just exceed them in some instances suggests that many low-paid workers are only just able to achieve the MIHL standard and are very vulnerable to even the slightest rise in living costs or decline in their incomes.

It is also clear that, as many others have observed, prevailing levels of NSA are woefully inadequate. The shortfall is AUD96 a week for a single person and varies for other family types between AUD47 for the sole parent with one child and AUD126 a week for the couple with two children. The AUD50 a week increase in the single rate of payment that has been advocated for by the Australian Council of Social Service (ACOSS, 2012) and other community groups would only reduce by about half the AUD96 a week payment shortfall for single adults. It would require a substantial increase in existing payment rates to close these adequacy shortfalls, which will continue to grow under existing policies.