3. Assumptions
3.1 The modelling considered four categories of expenditure:
- health expenditure (covering hospital, pharmaceuticals, Medicare and other, including the private health insurance rebate);
- income support (primarily the age pension, but also family benefits and a limited level of access to special benefits);
- expenditure on aged care services; and
- other expenditure (this related to the visa application administration costs, the Adult Migrant Education Program and the costs of administering the Assurance of Support).
3.2 Assumptions were required both in relation to what the current levels of Commonwealth expenditure are estimated to be and how these costs might be expected to grow over time. Assumptions were also required for mortality, levels of take-up and future trends in macroeconomic variables including inflation, wages and GDP (Gross Domestic Product). A full listing of the assumptions adopted is included at the end of this report. The following discussion deals with the derivation of the assumptions.
Health Costs
3.3 Health costs are the biggest contributor to total Commonwealth costs associated with the entrants under the CPV programs by a considerable margin. As has been well-documented by other agencies including AIHW and the Productivity Commission, health costs increase quite dramatically with age. Furthermore, the extension of procedures to the elderly that would previously have been restricted to younger age groups has seen health costs grow more quickly for the older age groups than for the population as a whole.
3.4 We have relied on two main sources of data in deriving the initial estimates of per capita health costs. The first is internal Treasury data showing average per capita expenditure for broad age groups. The figures cover Commonwealth expenditure on hospitals, the Pharmaceutical Benefits Scheme, Medicare and the Private Health Insurance Rebate in 2005/06. The second source was the Productivity Commission model built for their 2005 report on the Economic Implications of an Ageing Australia which included a distribution of per capita health expenditure by single year of age. These estimates were based on State and Commonwealth expenditure in 2002/03..
3.5 As we wanted per capita costs by single year of age, our approach was to assume that the single age relativities reported by the Productivity Commission were still applicable. We then rebased the Productivity Commission estimates to give the averages for the age bands reported in the Treasury numbers using population estimates as at 30 June 2006. The estimated per capita costs for 2005/06 were then inflated to take account of the expected growth in health costs over the two years to 2007/08.
3.6 The resulting estimates do not distinguish between the health costs for someone who is assumed to die during the year and someone who is assumed to survive. On average, the health costs in a year for someone who dies are considerably more than the costs for someone of the same age who does not. The exact multiple is open to question but a factor of 8 falls within the range generally considered reasonable and we have used this for our base scenario. The health cost for a survivor can then be estimated by taking account of mortality rates using the following formula:
where Px is the probability that an individual aged exactly x years survives to age x+1 years.
3.7 The health costs estimated for a survivor at each age are shown at Table 1 in listing of all assumptions following this report. A higher multiple will lead to lower estimated costs for survivors but the overall results are relatively insensitive to this variable.
3.8 Health costs have grown at rates well in excess of general growth in the economy over recent years and the view from both AIHW and the Productivity Commission is that this is likely to continue for some time into the future. Clearly at some point, this will become unsustainable but for the time periods involved in the current costing exercise we consider it reasonable to assume that the assumption recommended by AIHW of health cost inflation of 0.7 percentage points in excess of GDP growth will apply.
3.9 In order to take account of the recent experience of higher growth in health expenditures of older age groups, we looked at growth in hospital separations (that is, the number of people completing an episode as a hospital inpatient) as a proxy for the differential inflation. The following chart shows the percentage increase in hospital separations between 2001/02 and 2005/06 at different ages.
(Source: AIHW Australian Hospital Statistics 2005/06)
3.10 We have fitted a straight line to the public hospitals series to come up with a percentage increase at each age. Weighting these percentage increases by the population at each age gives an aggregate increase in health costs attributable to the changing hospital separation rates and a constant was added to this to give the overall increase of 0.7 percentage points above GDP.
3.11 It is not credible that these differential rates of increase will be maintained indefinitely into the future. At the same time, the considerable weight of people moving into the older age groups over the next few decades will lead to political pressure to increase resources devoted to these groups. We have used a logistic curve to moderate the difference from the long term average at each age so that after 20 years the rate of increase at all ages is 0.7 percentage points above GDP.
3.12 The initial difference from the long term assumption at each age is shown at Table 2 in the assumptions section at the end of this report.
3.13 An additional assumption was included in the model which allowed for the possibility that the groups under consideration are healthier than the general population. CPV applicants must undergo a health check prior to approval and they might therefore be expected to be somewhat healthier than the general population. This is taken into account by assuming that the health costs for a person of a particular age are taken to be the health costs for someone a specified number of years younger. Somewhat counter intuitively, the net effect of assuming a healthier population in this way is an increase in health costs. This is because of the excess health inflation and particularly the higher growth rates at older ages. In other words, because health costs are increasing faster than the discount rate used to calculate a present value of future outlays, the value of health expenditure for an eighty year old will be greater if the expenditure occurs later. Our base scenario assumes that visa populations will have the health characteristics of people three years younger. However, results are also included using the assumption that there is no difference between the groups of interest and the general population.
Income Support Benefits
3.14 Given the age distribution of the people covered by the CPV, the age pension is the income support benefit most likely to be paid. Under Centrelink rules, age pension is not payable until 10 years of residence have been completed. In the intervening period, it is possible that Special Benefits might be granted in hardship cases.
3.15 However, CPV applicants are covered by an Assurance of Support (AoS) arrangement which means that access to income support benefits prior to becoming eligible for the age pension is likely to be fairly minimal. In our 2002 report, we followed the then Department of Immigration, Multicultural and Indigenous Affairs assumption that 39% of applicants would have access to benefits before becoming eligible for the age pension. Discussions with the Department of Family, Housing, Community Services and Indigenous Affairs, and Centrelink regarding activation of the Assurance of Support suggest that this is substantially too high and for the current exercise we have assumed a take-up rate of 5%. We have assumed that this group will receive a benefit at a rate equal to 75% of the age pension that would otherwise be assumed to apply.
3.16 While the CPV demands significant resources in order to meet the second visa charge, there is no requirement on the financial resources which the applicant must have on entry. Accordingly, it is reasonable to assume that this group may be only marginally better off financially than the general Australian population in the same age group.
3.17 We used data from the Australian Bureau of Statistics (ABS) publication Household Income and Income Distribution Australia 2005-06 which showed the distribution of households by the proportion of their gross income which came from income support payments. For a given proportion, it is possible to calculate what the amount of pension must be based on the income test. Combining these pension amounts with the household distribution, an average pension entitlement across all households can be derived. This analysis was done separately for single and married couple households. In order to take account of the possibility that the CPV population is somewhat better off than the general population we adjusted the distribution slightly to overweight the proportion of households with low reliance on income support.
3.18 The following table sets out the underlying data and the distribution we used.
Single Households
| Percentage of gross income from income support | Percentage of households¹ | Adjusted percentage of households¹ | Implied annual pension entitlement² |
|---|---|---|---|
| <1% | 7.1% | 13.2% | $0.00 |
| ≥ 1%, < 20% | 7.8% | 14.5% | $3,389.04 |
| ≥ 20%, < 50% | 7.3% | 13.6% | $8,944.85 |
| ≥ 50%, < 90% | 21.1% | 21.1% | $13,308.20 |
| ≥ 90% | 56.5% | 37.7% | $15,268.16 |
Couple Households
| Percentage of gross income from income support | Percentage of households¹ | Adjusted percentage of households¹ | Implied annual pension entitlement² |
|---|---|---|---|
| <1% | 8.8% | 14.5% | $0.00 |
| ≥ 1%, < 20% | 13.3% | 21.9% | $5,688.35 |
| ≥ 20%, < 50% | 10.3% | 17.0% | $15,013.51 |
| ≥ 50%, < 90% | 25.8% | 25.8% | $22,337.17 |
| ≥ 90% | 41.7% | 20.9% | $25,626.89 |
- Note that column totals may not add exactly to 100% due to rounding
- Assuming the pension makes up a percentage of gross income at the middle of the reported band, calculated under current income test parameters
Source: Household Income and Income Distribution, Australia 2005-06 (ABS Publication 6523.0)
3.19 Applying the adjusted weightings to the implied pension entitlements gave an average pension entitlement for a single person of $10,260 per annum and $14,900 for a couple. We were able to link the CPV data to create records for each household unit and thus apply the appropriate rates based on the projected number of surviving adult members.
3.20 Reflecting the Government’s commitment to maintaining the single rate of age pension at 25 per cent of MTAWE (Male Total Average Weekly Earnings), pension rates are assumed to increase in line with wage growth.
3.21 There are also a small number of children in the CPV population, most of whom could be expected to qualify for Family Tax Benefit (FTB) while they are of an eligible age. We have made the following assumptions.
| Age of child | Percentage of children receiving Family Tax Benefit | Amount of Family Tax Benefit |
|---|---|---|
| 0 to 17 | 80% | $1,891 |
| 18 to 22 | 75% | $2,310 |
| 23 | 50% | $2,310 |
| 24 | 25% | $2,310 |
3.22 FTB is assumed to increase in line with the CPI (Consumer Price Index) rather than the wage inflation which applies to the age pension itself.
Aged Care Costs
3.23 Aged care expenditure through support for residential and community based services is the other major cost item. We have derived per capita aged care costs from the models developed by the Productivity Commission for their 2005 report on ageing. These models projected the costs to the Commonwealth of high and low care residential accommodation (nursing homes and hostels respectively), the Home and Community Care program and the Community Aged Care Packages together with usage rates of the different programs by age. We have converted this into an overall per capita cost by single year of age.
3.24 Aged care costs are driven primarily by wages and, accordingly, we have assumed that these costs will increase in line with wage growth. The initial per capita costs are shown in Table 5 in the assumptions section.
Other costs
3.25 Other costs arise from administration expenses, both direct and indirect, attributable to the AoS arrangements and the Adult Migrant Education Program.
3.26 Based on advice from DIAC, we have assumed that the visa processing costs will be $392.10 per application.
3.27 We were unable to get any advice on the administrative costs associated with the AoS. The advice provided by DIAC when we originally undertook this exercise suggested that the set up costs might have been of the order of $750 to $800 per application. We have assumed a cost of $800 per application in the first year and a further cost of $2,000 in the event that a claim is made for income support and the AoS is therefore activated. These costs are a very small proportion of the total.
3.28 DIAC advice suggested that around 27% of CPV applicants might access the Adult Migrant Education Program at a cost of $5,854 for each person spread over two years.
3.29 We have assumed that these other costs will increase in line with the CPI. While processing costs are expected to increase more slowly than this due to the efficiency dividend, it is not clear that this can be maintained in the long term. Again, the effect of this assumption is immaterial when compared with the other costs.
Mortality assumptions
3.30 We have modelled mortality based on the latest available Australian Life Tables: ALT2000-02. There is, however, a question around whether the health status of these two groups might be better than the general population and we have included a parameter which allows mortality to be adjusted to be equivalent to that of a person who is a certain number of years younger. Our base scenario assumes that applicants will experience the mortality of someone in the general population who is three years younger. As noted above, this factor is also assumed to flow through into health costs.
3.31 The mortality rates of the Australian population have improved substantially over the last century and it is quite likely that they will continue to improve. We have included a parameter which provides for continuing mortality improvement at either the rates observed over the last 25 or 100 years. The effect of including future mortality improvement is shown in the sensitivity analysis.
Financial assumptions
3.32 Assumptions are required on the rate of increase in the CPI, wage costs, GDP and a suitable discount rate.
3.33 We have maintained the assumption that the CPI will increase at a rate of 2.5% per annum. This is lower than the current rate of CPI growth but is in line with the Reserve Bank target range for price inflation and the Treasury long term view on inflation. This is also the rate that AGA currently uses in most of its long term projections. The rate of CPI growth is used to index the ‘Other Costs’ elements of the model.
3.34 The other financial assumptions have been set so as to be consistent with a CPI assumption of 2.5% per annum.
3.35 Wage costs are assumed to grow by 4% per annum. This again is consistent with the Treasury’s long tem outlook on the macroeconomy and is considered suitable for projections that extend 60 years into the future. The wage inflation assumption is used to index age pension entitlements and aged care costs.
3.36 Nominal GDP is assumed to grow by 5.5% per annum. GDP growth is used as the base to which the excess health cost inflation of 0.7 percentage points is added.
3.37 We have adopted a discount rate of 6% or 3.5 percentage points above the inflation assumption. This is in line with the assumptions we use for similar long term projections and is consistent with yields over the last few years on Commonwealth Government long term bonds.
3.38 It is probable that there will be net tax revenue arising from the CPV population. However, it is extremely uncertain. In our original costing, we restricted our analysis to the tangible costs under Commonwealth Government programs. Benefits, both tangible, in the form of provision of voluntary services or taxes which might be paid, and intangible in terms of the contribution of the migrants to Australian society more generally were ignored. We have maintained this approach for the current exercise.
Visa Charges
3.39 We have used the rates which applied at the time of preparing this report. For the CPV, we have used the distribution between the different visa types (onshore versus offshore, and single second visa application charge versus the split payment) shown on the actual intakes in each of the last three years. We have assumed that 5 percent of those who take up the option of a temporary visa will not choose to convert that temporary visa to a permanent visa.
Next: 4. The Model
Previous: 2. Data
Return to: Table of Contents