• No se han encontrado resultados

There were two analytical components to answering the question “how likely is it that a student will be able to afford a competitor school?” The first was to build an income profile for each SA1 in SEQ. The second was to estimate households willingness-to-pay for private schools, based on their income.

Building an income profile by SA1

Income data was collected at the SA1 level from ABS 2011 Census via TableBuilder.

The data were filtered by Family Composition (i.e. FMFC) at the 2-digit level.160 Families were excluded unless they were either a

159

See Coldron, et al. (2008). Also of note: using raw achievement as opposed to a measure of value add appears to be a good strategy for parents interested in maximising their child’s test scores, as illustrated in Allen and Burgess (2011).

‘couple family with children under 15’ or ‘one parent family with children under 15’.

The income measure used was Total Family Income (Weekly) (i.e. FINF).161 All 17 ABS FINF income bands were included. ‘Negative income’ was coded as having a value of 0 (in total, this band represented 0.6 per cent of the non-excluded families). Assessing the willingness to pay for private schools, by income group

Wave Ten of the Household, Income and Labour Dynamics in Australia (HILDA) was used to estimate the relationship between household income and the level of spending on education. The income variable used was jhifefp (household financial year gross income).162

No measure of ‘private school fees’ was available in the data. As an indicator, we used jhxyeduc (household annual expenditure - education fees).163 Four steps were taken to ensure that this variable best captured private school expenditure for secondary school students:

160

Australian Bureau of Statistics (2011b).

161

Australian Bureau of Statistics (2011c)

162

Melbourne Institute (2011). This variable was partly chosen based on a comparison of its density plot compared to the ABS measure FINF: i.e. this appeared to be the best analogue of the ABS variable.

163

Ibid.

• The combined HILDA sample (initial sample size of 17,855) was trimmed to exclude cases where household annual expenditure on education fees was < 0). This left 15,037 individuals in the sample.

• Data was confined to households with individuals between the ages of 12 and 17 (inclusive), which left 649 individuals remaining in the sample.

• The sample was filtered to exclude all individuals that attended government schools. This left 220 cases.

• Finally, the data were cleaned so that families with multiple children attending private school were only counted once. (The number of families with multiple children between 12- 17 attending private school was 12, bringing the final total to 208).164

This sample was then stratified into five income buckets. This was done to reflect the fact that within each income group, there will be a distribution of the willlingness to pay for private school fees. For every observed level of private school fees, the authors calculated the proportion of people in each income bucket who currently paid at least that figure in fees (based on the jhxyeduc variable). Figure 16 presents the plots for all five groups. It’s worth noting that in each group there are households who pay $0 fees

164

This means that in 12 cases, our affordability figures represent the cost of >1 student in a household. Given our fee data applies to a single student, the effect of this is to increase the estimated level of competition relative (by lowering the affordability barriers). That said, given the relatively small number of cases, this effect has no meaningful impact on our results.

(or very low fees). It is speculated that these students are on scholarships.

Combining income data in the sample area, and willingness to pay estimates

The income data, which originally had 17 bands, were rolled into the five income buckets defined in the ‘willingness to pay’ charts presented in Figure 16.

The proportion of people in each SA1 who would be able to pay was defined by the proportion of people in each income band multiplied by the proportion of people in that band who currently pay at least that much on private schooling.

Table 7 presents the example inputs for SA1 311303, and a school with Year 8 fees of $10,000. To estimate the likelihood that a family in this SA1 will be able to afford these fees, the sum- product of the two columns was calculated (which in this case yielded 29 per cent).

It’s important to note that this approach leads to a very generous estimate of ‘affordability’. Recall that the proportions defined by the charts in Figure 16 are based on the population of households who currently send their children to private schools. This is clearly not a representative sample of Australian households. This approach is adopted to be as generous as possible when assessing the current levels of competition between schools in Queensland.

Figure 16: Plot of actual fees paid (by household income group)

*Note: For this group, the x-axis is changed, as the sample included a number of families in the income group ‘$260,000+’ that paid more than $25,000 in education fees in 2010

Source: Melbourne Institute (2011) 0% 20% 40% 60% 80% 100% $0 $10,000 $20,000 0% 20% 40% 60% 80% 100% $0 $10,000 $20,000 0% 20% 40% 60% 80% 100% $0 $20,000 $40,000 0-$64,999 n=28 $65,000-$103,999 n=42 $104,000-$155,999 n=63 $160,000-$259,999 n=52 $260,000+* n=20 0% 20% 40% 60% 80% 100% $0 $10,000 $20,000 0% 20% 40% 60% 80% 100% $0 $10,000 $20,000

Fee data

Data on the fees being charged for Year 8 in 2013 came from three sources. In order of preference, these were:

- school websites, quoting 2013 rates (88 per cent of

schools). [Note that tuition charges were only recorded (i.e. not including extra charges such as capital levies,

application fees and so on). This, again, is a conservative approach in terms of access to private schools. In some cases, compulsory non-tuition charges make up nearly half of the total fees.]

- school websites, quoting 2012 or 2011 rates and converted into 2013 terms using the rough approach of applying the average annual increase in the education component of capital-city CPI increases from 2009-10 to 2011-12 (4 per cent of schools)165

- the My School website, based on ‘fees/student’ in 2011, put in 2013 terms (8 per cent of schools)166

165

Australian Bureau of Statistics (2012).

166

Using the inflator available from ibid.

Table 7: Example inputs for estimating the likelihood that a household in SA1 3100135 could afford fees of $10,000 p.a.

Income bucket Likelihood of people in income band to afford $10,000p.a. in

school fees (defined by distributions in Figure 16) Proportion of SA1 3100135 in income bucket 0 - $64,999 11% 31% $65,000 - $103,999 31% 23% $104,000 - $155,999 41% 38% $160,000-$259,999 38% 8% $260,000+ 80% 0%

Note: the slightly lower percentage of income group $160,000-$259,999 (relative to the $65,000-$103,999 group reflects the fact that in the HILDA wave 10, the observed level of private school costs for these groups was extremely similar)

Documento similar