Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Waiting lists to grow: 216,000 extra public hospital admissions expected.



Download PDFDownload PDF

M E D I A R E L E A S E

CONTACT: Kylie Walker CHA Communications Manager 02 6260 5980 / 0405 229 152

CHA is the largest single grouping of non-government health, aged and community care services in Australia.

12 June 2009

Waiting lists to grow: 216,000 extra public hospital admissions expected

Four times more people than predicted by Treasury may go without private health insurance as a result of changes to the private health rebate and Medicare levy surcharge, Catholic Health Australia (CHA) revealed today.

CHA, representing 75 Catholic not-for-profit hospitals, today published independent research that shows the Government’s new private health insurance changes could result in up to 100,000 people going without insurance.

The analysis, by Access Economics, found the rebate and Medicare levy surcharge changes could stretch public hospital waiting lists by an additional 36,000 people who would otherwise have been treated in private hospitals.

These forecasts come on top of Treasury predictions that 492,000 people are still expected to forgo insurance as a result of the 2008/09 Budget.

“When you combine the number of people likely to forgo their insurance as a result of this new change and last year’s changes, we will see 600,000 fewer people insured than would have otherwise been the case,” CHA CEO Martin Laverty said.

“This means some 216,000 extra admissions to public hospitals.”

The CHA research calls for extra funds to be provided to public hospitals operated by State and Territory Governments and the 21 public hospitals operated by Catholic Health Australia members, to cover the costs of treating 216,000 extra public patients.

It also indicates that the economic downturn is likely to prompt an additional decline in private health insurance coverage, which will result in even greater pressure on already stretched public services.

“We oversee 75 not-for-profit Catholic hospitals. Catholic hospitals treat both public and private patients. Our experience is that changes to private health insurance have their biggest impact on public patients,” Mr Laverty said.

“Treasury did not talk to health care providers before proposing these changes to private health. There is evidence Treasury barely consulted the Department of Health and Ageing until after these plans were well advanced.

“State, Territory, and Catholic Public Hospitals will need extra resources to meet this increase in demand, and we’re looking to the Federal Government to act.”

Commercial -in -Confidence

Impact of means-testing the PHI rebate and changing MLS parameters

June 2009

Report by Access Economics Pty Limited for

Catholic Health Australia

Impact of means-testing the PHI rebate and changing MLS parameters

Commercial -in -Confidence

Contents

Executive Summary and conclusions ..............................................................................................i

1 Introduction ........................................................................................................................ 1

2 Treasury’s modelling methodology .................................................................................... 3

3 Scenario analysis ................................................................................................................. 4

4 Comments on Treasury and PHI modelling methodology .................................................. 8

4.1 Treasury modelling in overview ............................................................................................. 8

4.2 The price elasticity assumption.............................................................................................. 8

4.3 Lack of stratification by policy type ..................................................................................... 10

4.4 Challenges in modelling PHI coverage ................................................................................. 10

4.5 Potential impact on public hospital attendances ................................................................ 11

Conclusions ................................................................................................................................. 13

Appendix A: Detail of the PHI budget measure ................................................................... 14

Charts

Chart 3.1 Net cost impact of savings measure on singles ............................................................. 6

Chart 3.2 Net cost impact of savings measure on families ........................................................... 6

Tables

Table 3.1 Preferences of policy holders by category .................................................................... 4

Table A.1 New PHI rebate parameters ........................................................................................ 14

Table A.2 Impact on cost of PHI to policy holder ........................................................................ 15

ii Commercial-in-Confidence

Glossary

CHA Catholic Health Australia

MLS Medicare Levy Surcharge

PHI Private Health Insurance

PHIAC Private Health Insurance Administration Council

Impact of means-testing the PHI rebate and changing MLS parameters

i Commercial -in -Confidence

Executive Summary and conclusions Access Economics was commissioned by Catholic Health Australia (CHA) to discuss the impact of means-testing the private health insurance (PHI) rebate and the related changes to the Medicare levy surcharge (MLS) announced in the 2009-10 Federal Budget. The focus is on the impact of the proposed policy on coverage (the number of Australians with PHI) and the subsequent impact on public hospital attendances. In addition, the veracity of the Treasury modelling methodology is discussed. The approach in this paper differs from other organisations that have used market research as the basis for projecting the impact of the proposed policy on coverage.

In brief, the report consists of:

■ a brief description of Treasury’s modelling of the PHI rebate budget measure; ■ the findings from our own scenario analysis; and ■ a discussion of the Treasury modelling and, more generally, the challenges that arise in modelling PHI coverage.

Treasury estimated that under these reforms, 99.7% of people will remain in private health insurance 1, a net loss of adults covered of only around 25,000. Some have argued that this will place extra pressure on the public hospital system as fewer people are covered by PHI. This report has examined the Treasury’s modelling methodology and found:

■ The Treasury’s personal income tax micro-simulation model is an appropriate tool for forecasting the impact of the means-testing of the PHI rebate and the associated changes in MLS thresholds and rates.

■ The forecast rests importantly on an assumption about the price elasticity of demand for PHI among those who are affected by the measure (higher income earners).

■ A proportion of PHI holders are likely to downgrade their insurance. The Treasury modelling did not incorporate the potentially important impact of this.

Given the short time period since the 2009-10 Federal Budget was presented to the Parliament, we have not undertaken comprehensive modelling of the Budget measure. We cannot, therefore, express a precise view on Treasury’s estimate that PHI coverage would be “around 25,000” adults lower (than it would otherwise be) as a result of the measure. On the basis of scenario analysis, we surmised that the measure would have its greatest impact in terms of loss of coverage on people in the lowest income tier, but that the “sharper sticks” (the higher MLS rates) would keep engaged those people in the two higher income tiers. In light of the range of potential plausible estimates of price elasticity of demand, we would not be prepared to rule out of contention an impact on coverage of minus 1% (approximately 100,000 people based on the current coverage of approximately 10 million people of all ages).

■ Based on population averages, the loss of 100,000 people would mean a 0.78% shift in inpatient workload (or a combination of extra workload and revenue loss equivalent to a 0.78% shift).

1 Treasurer’s joint press release with Nicola Roxon, The Minister for Health and Ageing, Rebalancing Support for Private Health Insurance, Canberra 12 May 2009.

ii Commercial-in-Confidence

In summary, we reached broadly the same conclusions regarding coverage as the Treasury based on our conceptually quite different scenario analysis. We do suspect, however, that some PHI fund members could respond to the measure by switching their cover to a lower feature, lower priced policy, something the Treasury model did not canvass. If that is the case, then the change in coverage per se may prove to be a flawed measure of the financial impact on the PHI funds and could lead to subsequent greater increases in premiums and further negative impact on membership levels and levels of cover held by those who retain their membership.

As a matter of judgment, we expect the recession to have a larger negative impact on coverage than the Budget measure. Likewise, people responding to tougher economic times may see a lower feature, lower priced policy as an option.

The broadening definition of MLS income will have a minor positive (thus offsetting) impact on PHI coverage as well as lifting MLS revenue.

Access Economics

Impact of means-testing the PHI rebate and changing MLS parameters

1 Commercial -in -Confidence

1 Introduction Access Economics was commissioned by Catholic Health Australia (CHA) to discuss the impact of means-testing the private health insurance (PHI) rebate and the related changes to the Medicare levy surcharge (MLS) announced in the 2009-10 Federal Budget. The focus will be the impact of the proposed policy on coverage (the number of Australians with PHI) and the subsequent impact on public hospital attendances. In addition, the veracity of the Treasury modelling methodology is discussed. The approach in this paper differs from other organisations that have used market research as the basis for projecting the impact of the proposed policy on coverage.

In brief, the report consists of the following sections:

■ a brief description of Treasury’s modelling of the PHI rebate budget measure; ■ the findings from our own scenario analysis; ■ a discussion of the Treasury modelling and, more generally, the challenges that arise in modelling PHI coverage; and

■ conclusions.

Description of MLS and PHI

The definition of income for the purposes of the MLS is changing. From 1 July 2008, the definition of income on which the MLS is based (‘MLS income’) was expanded to include reportable fringe benefit amounts and the amount on which family trust distribution tax has been paid. From 1 July 2009, the definition of MLS income will be expanded further to include the individual’s reportable superannuation contributions 2 and net investment losses 3. The widening of the MLS income definition will most likely lead to an increase in MLS revenue and

an increase in PHI coverage.

In the 2009-10 Federal Budget, the Government proposed as a savings measure to means-test the PHI rebate with a reduced rebate for two income tiers and no rebate for the highest tier of income earners. At the same time, the Government will progressively increase the MLS from 1% to 1.5% for those in the highest income bracket who do not have PHI. In general, those on

2 reportable superannuation contributions, for a person and an income year, means the sum of:

(a) the person’s *reportable employer superannuation contributions (if any) for the income year; and

(b) the total amount of contributions made by the person that the person has deducted or can deduct under Subdivision 290-C for the income year (Source: Tax Laws Amendment (Budget Measures No. 1) Bill 2009: Reforms to income tests). 3

total net investment loss of an individual for an income year means the sum of:

(a) the amount (if any) by which the individual’s deductions for the income year that are attributable to financial investments (within the meaning of the Social Security Act 1991) exceed the individual’s gross income for that year from those investments; and

(b) the amount (if any) by which the individual’s deductions for the income year that are attributable to rental property exceed the individual’s gross income for that year from rental property. (Source: Tax Laws Amendment (Budget Measures No. 1) Bill 2009: Reforms to income tests)

2 Commercial-in-Confidence

lower incomes will continue to be able to access the rebates at the current amounts and will remain exempt from the MLS. These changes will apply from 1 July 2010. A detailed description of the budget measure is included in Appendix A.

Impact of means-testing the PHI rebate and changing MLS parameters

3 Commercial -in -Confidence

2 Treasury’s modelling methodology The Treasury forecast of the effect of the measure on PHI coverage was generated using their personal income tax micro-simulation model. A micro-simulation model breaks down the “economic units” (sometimes individuals, in this case individuals and families) into very small sub-groups and predicts the response of each sub-group to policy change.

The most important assumption in the modelling was a price elasticity of -0.2. Price elasticity measures the responsiveness in the quantity demanded given a change in the price. A price elasticity of -0.2 means the demand for PHI is relatively inelastic, that is, it is relatively insensitive to price changes.

Treasury’s personal income tax micro-simulation model is based on a “confidentialised” sample of 2005-06 personal income tax data. Importantly, this data is not available to other stakeholders, so it is very difficult to replicate and second guess their work.

Other key items of data used in the Treasury modelling were the data on PHI membership as at 31 December 2008 4 and data on the average cost of a PHI policy. Treasury assumed that the average rate of PHI rebate is 32%. This average rate accords with the estimates of Access Economics.

In effect, Treasury modelled a change in the “net price” of PHI where that price takes account of the gross premium, the PHI rebate (if any) and the MLS surcharge that would otherwise apply if the taxpayer did not have a complying PHI policy.

Treasury concluded that the Budget measure would result in a net loss of adults covered of “around 25,000”. Some government documents give a more precise estimate of 22,500 adults. Treasury also found that those dropping (or not taking up) PHI as a result of the measure would come from the lowest income tier. They concluded that those in the top two income tiers would retain their membership.

The Treasury estimate cannot be read to mean that adult membership will decline by “about 25,000” from the current level (although it will be widely misunderstood to mean just that). It does mean that adult membership will be “about 25,000” lower than it would otherwise have been had there been no change to the PHI rebate. The Budget and forward estimates of PHI rebate expenditure implicitly rest on assumptions about levels of coverage, the age composition of the membership and changes in PHI premiums. However, the government does not reveal its base case assumptions. Indeed, it goes further than that. It secretes part of the

forward estimate for the PHI rebate in the contingency reserve to as not to signal to the PHI funds the future premium increases it expects to approve.

In short, because of the way the government holds its hands over the data, it is difficult to infer their expectations for the level of coverage. That said, we infer that the Government is expecting further rebate savings over time as a result of last year’s muted changes in MLS thresholds (the modified outcome of the savings measure announced in the 2008-09 Budget).

4 According to PHIAC, 9,702,117 Australians were covered for hospital treatment (44.6% of the population) in March 09, and in December 2008, there were 9,656,848 Australians covered for hospital treatment, (www.PHIAC.gov.au, accessed 7 June 2009)

4 Commercial-in-Confidence

3 Scenario analysis In the brief time period since the 2009-10 Federal Budget was presented to the Parliament, Access Economics has not undertaken comprehensive modelling of the Budget measure. However, we have undertaken some simple scenario analyses of the impact of the changes to the PHI rebate. We identified three broad types of PHI policies which, in turn, reflect our observations regarding the preferences of three broad groups of policy holders (see Table 3.1).

Table 3.1 Preferences of policy holders by category

Category of members Most likely type of policy

Group 1: Those with the highest incomes and/or the highest aversion to risk. Elderly single and widowed females are particularly risk averse. This group is generally assumed to have the lowest price elasticity of demand for PHI.

A high -end (higher priced) full y-featured product offering the most generous benefit entitlements.

Group 2: Those with middle incomes who see value in having PHI but who are also

concerned with the affordability of their cover. Also seek value. This group is generally assumed to be more sensitive to the price of PHI than group 1.

A low -premium product, most likely with the highest front-end deductible on offer and, if they are not at all risk averse, perhaps exclusions as well.

Group 3: The “surcharge avoiders” whose primary (if not sole) purpose in having PHI cover is to save money compared with any MLS surcharge they would otherwise pay. As Clayton’s members of a fund, they attach little or no value to any potential benefits they might be entitled to. Their decision to hold PHI (or not to hold as the case may be) might be influenced by the higher premiums that could apply under Lifetime Health Cover if they drop their cover and rejoin at a later date. These arrangements are, however, not well understood.

A low -premium product, most l ikely with the highest front-end deductible on offer and, if they are not at all risk averse, perhaps exclusions as well.

Source: Access Economics

The PHI funds do not share data on the composition of the policies they sell. Therefore, we are not able to quantify the numbers covered by each type of policy.

To further understand the impact of the PHI rebate and the MLS changes on each income group, Access Economics constructed scenarios for all three groups described in Table 3.1. The scenarios were based, respectively, on fully-featured PHI product with top level cover for hospital and ancillaries, a “value” product with medium-range benefits and a front-end deductible, and a very basic product that would appeal to a person wishing to avoid paying the

Impact of means-testing the PHI rebate and changing MLS parameters

5 Commercial -in -Confidence

MLS. A fourth scenario was constructed for the “average” policy (inferred from data published by the Private Health Insurance Administration Council (PHIAC)).

The financial impact of the measure on singles and families is depicted in Chart 3.1 and Chart 3.2 respectively. The net cost of health insurance is defined as the cost of the policy (net of the rebate, if any) less the cost of the MLS that would apply if the person or family did not have a complying PHI policy. The charts illustrate the dollar change in the net cost.

The scenario analysis suggests the budget measure has its greatest impact on the lowest income tier, those with incomes for MLS purposes of $75,000-$90,000 (singles) or $150,000 to $180,000 (families). People in these income groups face increases in the net cost of between 25% and 40%. Some within this tier would no longer find it beneficial to retain or purchase PHI

once the Budget measure is implemented.

People in the middle income tier can face even larger percentage changes in the net cost. However, PHI is community rated (premiums are product-based so all who buy a particular product pay the same gross price regardless of the level of their income). The measure is less likely to prompt those on middle incomes to drop their cover because PHI premiums represent a smaller proportion of their incomes.

For people in the top income tier, the net cost of cover generally declines as a result of the measure. These people are the least likely to drop their cover. Given the expanded definition of income for MLS purposes, a small number who now choose to pay the MLS may take up PHI. They are clearly better off financially if they take up or retain a low-cost policy even if they attach no other value to the insurance in its own right.

The measure applies no more “stick” to the lowest income tier but a lot more “stick” to the highest income tier.

The scenario analysis also shows that the surcharge avoiders will, for the most part, find it worthwhile continuing to hold a low cost policy.

Not all people in high income categories will feel well off in an absolute sense. Some would have been particularly affected by the global financial crisis. At the same time, they are the group most likely to have PHI. Along with those who are highly risk averse, the price of PHI is rarely a compelling consideration. The size of the increases in the net cost, especially in the case of a top level product, may, however, prompt members to review the level of their cover. They may choose to moderate the cost impact by switching to a product with less generous benefits, by dropping or reducing their cover for ancillaries or by electing a larger front-end deductible amount.

As noted earlier, the Treasury model was based on an “average” premium. It is, therefore, not suitable for assessing the extent to which those with PHI downgrade their cover in response to the measure. Downgrading also reduces rebate expenditure but it will not cause as much collateral damage to the public hospital system (through the transfer of in-hospital episodes). Downgrading transfers responsibility for payment for some services from insurers to patients, with potential consequences for hospital transaction costs.

6 Commercial-in-Confidence

Chart 3.1 Net cost impact of savings measure on singles

-$500

-$400

-$300

-$200

-$100

$0

$100

$200

$300

$400

$75,001 $80,001 $85,001 $90,001 $95,001 $100,001 $105,001 $110,001 $115,001 $120,001 $125,001 $130,001 $135,001 $140,001 $145,001 $150,001

Change in net cost of PHI

MLS income higher than …

Value policy

Average policy

MLS avoider

Top Level

Source: Access Economics

Chart 3.2 Net cost impact of savings measure on families

-$1,000

-$800

-$600

-$400

-$200

$0

$200

$400

$600

$800

$150,001 $155,001 $160,001 $165,001 $170,001 $175,001 $180,001 $185,001 $190,001 $195,001 $200,001 $205,001 $210,001 $215,001 $220,001 $225,001 $230,001 $235,001 $240,001 $245,001 $250,001 $255,001 $260,001 $265,001 $270,001 $275,001

Change in net cost of PHI

MLS income higher than …

Value policy

Average policy

MLS avoider

Top Level

Source: Access Economics

Impact of means-testing the PHI rebate and changing MLS parameters

7 Commercial -in -Confidence

People who currently hold PHI may respond to the Budget measures by:

■ paying the extra cost and keeping their existing policy; ■ dropping their cover and taking their luck with health events and public hospital waits; or

■ downgrading their existing policy by switching their cover to a lower feature, lower priced policy.

Access Economics speculates that Lifetime Health Cover may prompt some price-sensitive members to elect the third option rather than the second. This is because the minimum requirement to avoid the MLS is to have an insurance policy that includes hospital cover (whether by itself or combined with extras and ancillaries). Therefore, price-sensitive members will most likely downgrade their policies by dropping their ancillaries and extras but maintain hospital cover so as not to attract the MLS. Price insensitive members will gravitate to the first.

8 Commercial-in-Confidence

4 Comments on Treasury and PHI modelling methodology

4.1 Treasury modelling in overview It is generally accepted that micro-simulation models are the best tools for analysing the effects of policy changes of this nature. That said, the quality of the modelling depends in part on whether the model can be fed the appropriate cohorts (this is addressed in section 4.3). It also depends on the extent to which the data sets underpinning the model are current or dated.

The Treasury model uses a “confidentialised” sample 2005-06 personal income tax data. The quality of the sample is not known, so the extent of sampling error cannot be estimated. In 2005-06, MLS liability related to taxable income. It now relates to a modified concept (referred to above as MLS income), importantly including reportable fringe benefits and reportable superannuation contributions. Accordingly, the 2005-06 data may not give an accurate picture of the number of individuals or families in each of the new MLS income tiers.

It appears that the Treasury has somewhat blithely assumed that those making salary sacrifice superannuation contributions will all be in the higher MLS income tiers anyway.

4.2 The price elasticity assumption Treasury has noted that academic and empirical evidence suggests that the price elasticity of demand for PHI is around -0.3 for all income categories. However, the Budget measures affect higher income groups only as lower income groups maintain their status quo. Accordingly, the Treasury used a lower price elasticity assumption of -0.2 because those with higher incomes are less likely to demand lower levels of PHI because of an increase in price compared to lower income groups. The Treasury has acknowledged that this comes down to a matter of judgment. We are not aware of any specific studies of higher income earners’ price elasticity of demand for PHI.

Price elasticity is a measure of the responsiveness of the quantity demanded of a good or service to any change in the price. A price elasticity parameter of -0.2 implies that the demand is relatively insensitive to the price. A simplistic way to express it, in mathematical terms, is that the price elasticity of demand is the percentage change in volume divided by the percentage change in price. If the price rises by 10% and that causes the volume of sales to fall by 2%, then the price elasticity of demand is -0.2 (-2% divided by 10%).

There is a more complex concept called point elasticity. To illustrate this, suppose that a good or service is priced at $100 and has an elasticity of demand estimated at -0.2. From that $100 starting point, it is possible to estimate the effect on volumes of moving the price up or down. It is possible that the price elasticity of demand is not constant at all points along the demand curve, and there is a range of prices at which demand becomes more responsive to changes in price. For example, if the price was changed to $200, then the price elasticity of demand at

that point might turn out to be higher than -0.2, say -0.5. A higher elasticity of -0.5 will lead to much larger potential drop outs or policy downgrades, given the large percentage changes in the cost of PHI net of the new rebate faced by some groups. We did not have access to the data for any detailed modelling exercise.

Impact of means-testing the PHI rebate and changing MLS parameters

9 Commercial -in -Confidence

Surveys of consumer attitudes find that the price of PHI is a relatively low-ranking consideration among those who take it up. Adverse selection means that the most risk-averse people will be most likely to take up or retain cover (due to information asymmetry between those who hold and sell insurance policies).

It is known that people on higher incomes:

■ have a much higher take-up of PHI (as they are more able to afford it); ■ have a higher opportunity cost (in terms of income foregone) when they have to take time off work for health care;

■ are generally better-informed and more able to access and assess health information than people on low incomes (it is relevant that there is a strong correlation between education attainment and income levels);

■ are more likely, when they need a health procedure, to want to exercise choice (or at least have some say) as to what is done (which procedure), when it is done (given that PHI is, in essence, queue-jumping cover), where it is done (which hospital or day surgery facility) and who does it (which doctor); and

■ have an extra (negative) incentive in the form of the MLS.

We conclude that it is intuitively correct that the higher income earners will have a lower price elasticity of demand for PHI than the population more generally. Thus, Treasury’s assumption does not appear unreasonable. It is, however, a reflection of a judgment as opposed to being based on strong empirical evidence. It should therefore be regarded as being within the range of plausible assumptions. It would be more informative, of course, to model the impact using a range of reasonable price elasticities to obtain a minimum and maximum net cost effect of the Budget measures (ie. to undertake probabilistic modelling).

Treasury’s estimate of the impact of the measure is variously stated as a loss of 22,500 adult members or 25,000 adult members. Based on the overall profile of PHI membership, a loss of 25,000 adult members would imply a fall in coverage of 33,000 persons of all ages.

Our own scenario analysis leads us to suspect that the Treasury estimate may be at the lower end of the range of possibilities. We would not rule out a fall in coverage of 1% of the membership (100,000 people). In other words, coverage could decline by 100,000 from where it would otherwise have been. This is not the same thing as saying that coverage will fall by 100,000 from the current level.

Our reasons for arguing for a larger reduction in coverage arise because the net price impact is largest for those in the lowest of the three income tiers (they are also the most numerous group). Given their lower incomes, we would expect them to have a higher price elasticity of demand than the higher income earners in the top two tiers. The Treasury model treats all people in all three tiers as having the same price elasticity of demand for PHI.

We stress that in the limited time available since the measure was announced, we have not been able to construct a detailed econometric model.

10 Commercial-in-Confidence

4.3 Lack of stratification by policy type As noted in Section 3, Access Economics undertook scenario analysis based on three main groups of PHI policyholders and their expected type of cover. In response to our questions, Treasury indicated that they had stratified by income alone.

In an ideal world, the stratification would have been by both income and type of policy held. Treasury noted that the various data sets are disconnected and that this obstructs the more complex cross-classification. While this is a valid defence, it still remains the case that the model rests on the price effects for the average premium, whereas holders buy a range of policy types and the price bands are quite wide. Another consequence of a model based on an average premium is that it really does not envisage that PHI fund members could respond to the measure by switching their cover to a lower feature, lower priced policy (a policy they deem to be “affordable” in current personal or economy-wide circumstances).

4.4 Challenges in modelling PHI coverage PHI is a complex product, one that is not well understood by consumers. Based on the complaints lodged with the PHI Ombudsman, there is a considerable degree of confusion on issues such as the expected benefits, the caps on entitlements and the gaps between fees and benefits. There is also evidence that the government incentive programs are not well known and well understood. We expect, for example, that very few people were aware that the definition of income for MLS purposes was widened in 2008 to include reportable fringe benefits and the amount on which family trust distribution tax has been paid. Indeed, the web sites of some PHI funds continue even now to describe the MLS income test with reference to taxable income alone. In a similar vein, we expect that even fewer people are aware that the definition of income for MLS purposes will be widened further to encompass reportable superannuation contributions and total net investment losses.

Given the complexity of the product and the extent of consumer ignorance (and indifference) to the fine detail, it cannot be expected that all householders will act in a manner that might be argued to be in their best financial interests. To illustrate that point, a number of taxpayers choose to pay the MLS when they could save money by taking out a basic (low feature, low price) PHI policy. Some people have an ideological objection to PHI. Others have had to pay the MLS simply because they did not see it coming.

The micro-simulation modelling implicitly assumes that people react to price (to price alone, actually) and, indeed, that they are able to accurately calculate the net price. Where people are unable to do so, their behaviour may depart from the model predictions.

The measure is estimated to improve the budget bottom line by $1.9 billion (over 4 years). It is of interest that Treasury estimated that some $150 million of that total is attributable to a revenue effect (even more people paying the MLS, or some people paying a higher MLS which they could avoid, while saving money, by taking out low-end cover).

It is clear from the history of PHI that factors other than the price of PHI can, from time to time, have a material impact in take-up. For example:

■ The 1984 NSW doctors’ dispute visibly affected public confidence in the public hospitals and resulted in a rise in coverage in NSW in particular. More recent high profile events

Impact of means-testing the PHI rebate and changing MLS parameters

11 Commercial -in -Confidence

(deaths by hospital care misadventure, miscarriages in hospital toilets) would likewise have had an impact on confidence in the public hospitals; and

■ PHI is commonly perceived as queue-jumping insurance which gives access to elective surgery (cf. long waiting times for access as a public patient). Government expenditure on elective surgery in public hospitals can change perceptions of both the capacity of the public system and the need to retain PHI cover.

It is a rare event when only one explanatory variable changes in a period of time (thereby allowing the impact to be observed). In the year ahead, we would expect the following factors to have some impact on PHI coverage:

■ the measure examined in this report (the means-testing of the PHI rebate and the related changes in MLS thresholds and rates);

■ the broadening in the definition of income for MLS purposes; ■ some residual effects from the changes in the MLS income thresholds in the 2008-09 Federal Budget;

■ the changes in MBS rebates for in-hospital items (cataracts and varicose veins rebates down, delivery rebates up) which will result in extra costs in some cases, or cost savings in others, if PHI gap schedule fees do not change. This is a “damned if you do, damned if you don’t” proposition for the PHI funds. If they try to pass through MBS cuts, it will damage the reputation of the gap insurance products. If they absorb the MBS cuts, they’ll likely have to increase their premiums to cover the extra cost of benefits paid;

■ the ongoing maladministration of public hospitals by the State and Territory governments;

■ the deleterious consequences for the public hospitals of the economic slowdown induced loss of GST revenue by State and Territory governments; and

■ the impact of the economic slowdown on the affordability of PHI. In short, there will a great deal of “noise”. When we are able to look back at these years, with hindsight, we will not find it at all easy to divine how much each of these factors would have contributed to the overall outcome. As a forecasting proposition, it is even harder to isolate and forecast the impact of just one of these factors in the future.

Furthermore, as noted above, the nature of the forecast itself is often misunderstood. Treasury is predicting that, as a result of the Budget measure, the adult coverage of PHI will be “around 25,000” lower than it would otherwise have been . This is not the same thing as saying that adult coverage will fall by 25,000 from the current level. The baseline forecast (a forecast of coverage in the absence of the measure) has not been revealed.

4.5 Potential impact on public hospital attendances The subsequent potential impact of public hospital attendances is quite difficult to assess. Adverse selection would indicate that those most likely to drop their private health cover in response to the measure will be those who least expect to need it. Of course, nobody knows for sure when they will need access to a hospital. Young men might be relatively free of degenerative conditions, but they do have more accidents.

12 Commercial-in-Confidence

If the “drop outs” are concentrated among the “surcharge avoiders”, then there will be little or no effect on public hospitals. The surcharge avoiders are already using them.

If, perchance, the “drop outs” are concentrated among the older, high service using cohorts (say self-funded retirees in the bottom tier, for whom the PHI price change is the last straw), the impact will be more significant.

Based on the most recent Australian Institute of Health and Welfare hospital data, 100,000 people would account for approximately 37,000 5 separations in 2007-08 which, if transferred to the public sector, would represent an increase of just over 0.78%. The impact on accident and emergency (A&E) services might be relatively larger given that private A&E services are

limited.

Treasury chose to express their estimates in terms of the number of adults. This leads to a lower number than when expressed as people covered. Based on a comparison of PHIAC data for numbers of policies as opposed to numbers of people covered, 25,000 adults equates to approximately 33,000 people dropping their cover. The separation rate needs to be applied to people covered, not adults. If Treasury's 492,000 people covered from the 2008-09 Budget measure last year is added to the 25,000 adults from the most recent 2009-10 Budget measure, we estimate that 525,000 people of all ages will no longer be covered by PHI. This is equivalent to over 196,000 extra separations shifted to the public system (+4.1%). According to the Access Economics scenario (that 100,000 people drop their coverage in response to the 2009-10 Budget measure), 592,000 people would no longer be covered. In this case, there would be nearly 220,000 extra separations shifted to the public system (+4.6%).

Of course, if the impact on coverage is as low as the 25,000 adults forecast by the Treasury, and previous policy changes are not factored into the analysis, then the impact on public hospital separations is more modest, in broad terms a 0.3% impact (on average) on

separations and a bit more again on A&E services.

This would be a material increase, especially in the current fiscal environment where State and Territory governments are struggling to cope with a significant slowing, for a time, in the growth of GST revenues. It is noted that some privately insured patients now use the public

system (as private patients). If they drop their cover, their status changes to public patient and the effect on the public hospitals is loss of revenue rather than extra cost associated with extra volume.

A case can be made for provision in AHCA funding for compensation payments to State and Territory governments when, as a result of changes in the level of Federal Government support for PHI, workload is shifted to the public sector or revenue is lost when patients’ status changes.

5 According to AIHW Australian Hospital Statistics 2007-08 (Cat. No. HSE 55) released 10 June 2009, there were 7.874 million separations in 2007-08, a separation rate of 0.3722 per person.

Impact of means-testing the PHI rebate and changing MLS parameters

13 Commercial -in -Confidence

Conclusions As noted earlier, in the brief time period since the 2009-10 Federal Budget was presented to the Parliament, we have not undertaken comprehensive modelling of the Budget measure. We cannot, therefore, express a precise view on Treasury’s estimate that PHI coverage would be “around 25,000” adults lower (than it would otherwise be) as a result of the measure. On the basis of the scenario analysis (also described in Section 3), we surmised that the measure would have its greatest impact in terms of loss of coverage on people in the lowest income tier, but that the “sharper sticks” (the higher MLS rates) would keep engaged those people in the two higher income tiers. Given the range of potential plausible estimates of price elasticity of demand, we would not be prepared to rule out of contention an impact on coverage of minus 1% (approximately 100,000 people based on the current coverage of approximately 10 million people of all ages).

Based on population averages, the loss of 100,000 people would mean a 0.78% shift in inpatient workload (or a combination of extra workload and revenue loss equivalent to a 0.78% shift).

In summary, we reached broadly the same conclusions regarding coverage as the Treasury based on our conceptually quite different scenario analysis. We do suspect, however, that some PHI fund members could respond to the measure by switching their cover to a lower

feature, lower priced policy, something the Treasury model did not canvass. If that is the case, then the change in coverage per se may prove to be a flawed measure of the financial impact on the PHI funds and could lead to subsequent greater increases in premiums and further

negative impact on membership levels and levels of cover held by those who retain their membership.

As a matter of judgment, we expect the recession to have a larger negative impact on coverage than the Budget measure. Likewise, people responding to tougher economic times may see a lower feature, lower priced policy as an option.

The broadening definition of MLS income will have a minor positive (thus offsetting) impact on PHI coverage as well as lifting MLS revenue.

Access Economics

14 Commercial-in-Confidence

Appendix A: Detail of the PHI budget measure Effective 1 July 2010, the Government intends to means-test the PHI rebate and to introduce higher MLS rates for higher income earners who do not have a complying PHI policy.

The changes relate to three tiers of income. People in the lowest tier of income will face the same MLS rate as now (1.00%) but will receive a lower PHI rebate rate. People in the middle tier of incomes will face a higher MLS rate (1.25%) but will receive an even lower PHI rebate rate given a sliding scale. People in the top tier of incomes will face the highest MLS rate

(1.50%) but will receive no PHI rebate at all.

Singles with income of up to $75,000 and families with income of up to $150,000 will not be subject to any changes as part of this measure. They will continue to receive a PHI rebate, depending upon their age, of 30%, 35% or 40%. Singles and families with incomes below the current thresholds for the MLS will continue to be exempt from it. The government envisages that the current MLS thresholds will, as a result of indexation, align with the thresholds in this measure by 2010.

The new parameters for the rebate and the MLS are summarised in the following table.

Table A.1 New PHI rebate parameters

Income PHI rebate Medicare levy surcharge

Singles

Less than $75,000 No change No change

$75,001 - $90,000 20% (to age 65) 1.00%

$75,001 - $90,000 25% (age 65-69) 1.00%

$75,001 - $90,000 30% (age 70+) 1.00%

$90,001 - $120,000 10% (to age 65) 1.25%

$90,001 - $120,000 15% (age 65-69) 1.25%

$90,001 - $120,000 20% (age 70+) 1.25%

$120,001 & over Nil 1.50%

Couples

Less than $150,000 No change No change

$150,001 - $180,000 20% (to age 65) 1.00%

$150,001 - $180,000 25% (age 65-69) 1.00%

$150,001 - $180,000 30% (age 70+) 1.00%

$180,001 - $240,000 10% (to age 65) 1.25%

$180,001 - $240,000 15% (age 65-69) 1.25%

$180,001 - $240,000 20% (age 70+) 1.25%

$240,001 & over Nil 1.50%

Source: Budget papers

Impact of means-testing the PHI rebate and changing MLS parameters

15 Commercial -in -Confidence

Income in this context refers to income for MLS purposes. The definition of income for MLS purposes changed from 1 July 2008 and will change again from 1 July 2009.

The government intends that all income thresholds will continue to be indexed to wages with the stated objective of keeping these changes fair and sustainable into the future.

The government estimates that the measure will yield net savings of $1.9 billion comprising a reduction in the private health insurance rebate of $1.8 billion over four years and an increase in revenue through the surcharge of $145 million over the same period.

The total cost to implement this measure is estimated at $69.0 million over five years, which includes $66.6 million for the Australian Taxation Office, $1.9 million for the Department of Health and Ageing, and $540,000 for Medicare Australia.

The stated justification for the measure is that private health insurance is an area that departs from the general principle of Australia’s tax and transfer system that the largest benefits are provided to those on lower incomes.

In broad terms, the measure potentially affects some 2 million covered by PHI but another 8 million low and middle income earners will see no change.

The following table shows that the net (out-of-pocket) cost of PHI to the policy holder after taking account of the reduced/removed PHI rebate will rise from between 14.3% (lowest tier) and 42.9% (highest tier) while the cost of the MLS for high income earners without PHI will not change for the lowest income tier, will rise 25% for the middle income tier and 50% for the top income tier.

Table A.2 Impact on cost of PHI to policy holder

MLS income range

% increase in out-of-pocket PHI cost % increase in out-of-pocket MLS cost

$75,000 to $90,000 (singles) $150,000 to $180,000 (couples) 14.3% 0%

$90,000 to $120,000 (singles) $180,000 to $240,000 (couples) 28.6% 25%

$120,000 and over (singles) $240,000 and over (couples) 42.9% 50%

Source: Access Economics