Number 16: Cost–benefit analysis of portability policy
This report was published by the former Department of Families, Community Services (FaCS).
- Executive summary
- 1. Overview of portability policy and cost-benefit methodology
- 2. Social benefit of portability policy
- 3. Social cost of long-term portability of age pensions
- 4. Conclusion—net cost-benefit of portability policy
- Appendix 1
Definition, purpose and method
The term 'portability' in our context refers to the continuation of social security payments during a recipient's absence from Australia. Portability of Australian social security payments has always been a contentious issue for at least two reasons. First, the Australian social security system is based on residence and need. To be entitled to payments people are not required to have made contributions or paid taxes. Therefore, long-term or permanent portability may be considered inconsistent with a residence-based system. Second, long-term or permanent portability may be viewed as costly and, in providing payments to people who cease to be Australian residents, unfair to the Australian taxpayer.
The purpose of this paper is to examine the social and fiscal effectiveness of portability policy. Based on analysis carried out between November 1998 and November 1999, it is the first comprehensive assessment of portability policy since its introduction in 1972. To assess the social benefits produced by portability policy and its costs, including the costs of pensions as well as some external factors such as the costs of welfare infrastructure, a standard cost-benefit method was applied.
The paper first presents an overview of portability policy. The forces of migration, globalisation and other phenomena of the late twentieth century have had a role in shaping the development of Australian portability policy. For example, the overseas rate of Australia payments has become related to the length of Australian working life residence. This change reflects the relationship between the period of contributions and payment rate administered by other countries.
Most Australian social security benefits can be paid to customers who travel overseas for a short period of time. However, to test the social benefit of portability policy, the paper focuses on the portability of Age Pension. There are two major reasons for this focus. First, among the Australian community, Age Pension is probably the best-known social security payment. It is also associated with a part of life that many people look forward to and plan activities such as travelling overseas. Testing the attitudes of the general public to the portability of Age Pension will therefore produce meaningful results. The second reason for this focus is that Age Pension can be portable indefinitely and many Australian age pensioners already reside overseas. This allows for calculating the costs and savings of portability policy. The availability of historical data on Age Pension portability also allows for the identification of trends and the modeling of possible future changes.
Social benefits and costs
Chapters 2 and 3 are devoted to the social benefit and costs of portability policy. The best evaluation of the efficient use of financial resources is through an estimation of the social benefit produced by the policy. The social benefit of portability policy is calculated on the basis of a specially designed survey, popularly known as a contingent value survey. Delivery of the survey was contracted out to Roy Morgan Research. Respondents from 1 009 randomly selected households were asked three sets of questions. The first set identified the demographic characteristics of the respondents. The second set asked general questions about lifestyle, attitudes and plans for retirement. The last set of questions was designed to obtain an evaluation of portability policy in monetary terms. Respondents were asked to imagine themselves in a situation in which they have retired and are receiving Age Pension. This pension would not be payable during any overseas absence unless they are prepared to pay a small fortnightly fee, similar to an insurance premium. Respondents were initially asked whether they were prepared to pay a certain defined fee. To verify the accuracy of their valuations and to provide them with an opportunity to reconsider their previous decisions, respondents were also asked a question about the maximum fee they would be prepared to pay.
The survey showed that more than 81 per cent of respondents highly value short-term portability of less than 6 months. This support is equal across all respondent groups. A smaller proportion of respondents, 23 per cent, highly value long-term or permanent portability of Age Pension. The survey also shows that, were pensions to cease on departure, respondents intending to go overseas permanently are more likely to go than those intending to go overseas temporarily. However, the overall percentage of people willing to go overseas regardless of the portability of their pensions was small.
Two amounts were considered in the calculation of the social benefit of portability policy: the amount declared on the basis of the fixed value questions, where the initial $2 fee was accepted by 76.5 per cent of respondents, and the amount delivered on the basis of the open fee question. The total amount delivered on the basis of the survey questions was weighted and multiplied by a coefficient of Australian population. The monetary value delivered in this way reflected the total social benefit. However, it has to be stressed that it had no real monetary value and it was only used for the purpose of comparing, on equal terms, the costs of portability policy with the social benefit.
The total cost of portability policy was calculated by taking into account the cost of pensions paid overseas to people who would have left regardless of portability and the benefits forgone such as taxes, earnings on capital taken overseas, child care and other services provided by pensioners. Savings from portability policy were calculated by considering the costs of community infrastructure, such as health care and welfare services.
Net cost was calculated under three scenarios based on the percentage of people who would have gone overseas regardless of portability policy—that is, if 14 per cent, 19.25 per cent or 24.5 per cent of pensioners go long-term overseas, regardless of portability of their pensions. The overall result was that, under these scenarios, there are savings ranging from $17 million to $39 million a year. This contrasts with people who intend to travel short-term overseas. They would not go overseas if their pensions were not portable and short-term portability has no impact on costs of pensions and welfare infrastructure. Therefore, short-term portability policy appeared to be cost-neutral.
Conclusions and recommendations
Having considered both the social benefit value expressed in monetary terms and the total savings produced by portability policy, the paper concludes that portability policy is of great benefit to Australian society. Not only are savings made, but the average social benefit is also nearly five times higher than savings produced by the policy. Even if we consider some of the shortcomings of the cost-benefit method, overall portability produces a very high net social benefit. If it is considered that the sampling variation is between +/- 3 per cent, the magnitude of the overwhelming benefit would barely change.
The paper also recommends that, because short-term portability produces the highest social benefit, only policy related to long-term portability may be adjusted further. Some consideration may be given to technical adjustments designed to increase the positive financial impact of long-term portability. Also, in the context of the criticism of portability policy sometimes seen in the mass media, raising the level of awareness in the community of the social and financial benefits of portability policy may be a worthwhile investment.
- 1.1 Purpose of the research
- 1.2 Development of Australian portability policy
- 1.3 Application of cost-benefit methodology to portability study
- 1.4 Stages of cost-benefit analysis of portability policy
The term 'portability' refers to the continuation of social security payments during a recipient's absence from Australia. The portability provisions of social security law set out which payments are portable, the periods for which they are portable and the rate payable during any period of portability.
Portability is the cause of some 'systemic' friction within the Australian social security system. Contributory social security systems, which are found worldwide, acknowledge the acquired right of an individual to benefits. This right is derived from the financial contributions made during a person's working life. Portability is an extension of that right. Most contributory systems, particularly those operating in the member states of the European Union, have for many years encouraged the portability of benefits for the major contingencies of life: age, invalidity and widowhood. However, some of these systems link portability to social security agreements (for example, Belgium and Switzerland) or to reciprocal portability rights (for example, the United States).
The Australian social security system is based on residence and pays benefits according to need. Qualification for payments is dependent upon a claimant having spent a minimum period as an Australian resident. A claimant must also be an Australian resident at the time of the claim and/or while the payment is current. Also, before a payment may be granted, a claimant must satisfy the asset and income tests. The portability of benefits may be seen as compromising this policy objective and as creating particular administrative and control problems.
Nevertheless, in line with international practice, Australia has made many of its social security benefits portable, beginning with limited portability under its long-standing agreements with New Zealand (1949) and the United Kingdom (1953), and progressing to indefinite portability under new portability legislation in 1973.
During the 1972 parliamentary debate on the payment overseas of Australian social security pensions, two major issues were discussed. In principle, the Government (the Coalition) and the Opposition (the Australian Labor Party) agreed that,'if people have earned social service benefits in Australia, it should be no business of the government where those people choose to receive their benefits'.1 The central issue was how long people should be required to have resided in Australia in order to be entitled to overseas payment of their pensions. Arguments also revolved around how much tax a resident should had to have paid before being able to choose to leave the country. In the context of this debate, the right to portability of Australian pensions was linked to the period of tax contributions.
More recently, the arguments for and against the portability of Australian pensions have been concentrated around the issues of reciprocity, globalisation and human rights. Portability is considered good international practice in a globalised world. The introduction of a proportional rate of Australian pension for people with less than 25 years' working-life residence was not done in relation to the payment of taxes. The major concern during the parliamentary debate was that payment of Australian pensions overseas should reflect international practice.2
From the very beginning, portability policy has also been discussed in the context of Australian social security agreements. As a unilateral initiative to pay Australian pensions overseas, portability policy was considered too generous. However, quite often during parliamentary debate the unique character of the Australian social security system also emerged as a major obstacle to the negotiation of new agreements. Ideally, Australia would have a network of shared-responsibility agreements with all migrant source countries. Since this outcome was not seen as practicable, portability policy has usually been supported because it is considered to be a right of Australian residents.
Another context for portability policy is the diverse composition of Australian society. Having spent their working life in Australia, many migrants want to retire to their country of origin. Australian-born people are also increasingly deciding to live overseas. But portability policy is often incorrectly seen as designed to satisfy the needs of the migrant community and to be of little value to 'mainstream society'.
The major purpose of the research was to learn about the social benefit of portability policy and its fiscal implications. To address the concerns outlined above, we focused on portability of Age Pension and explored the attitudes of the general Australian public. For methodological purposes we accorded portability of Age Pension the status of a right derived from a person's qualification for the payment. This is a simplified portability model created for the purpose of this research project only.
The research investigated the problem of whether portability policy produces a benefit for Australian society and whether or not the general public appreciates long-term portability. It also throws some light on whether the payment of pensions to customers overseas on a long-term basis generates additional costs or savings. These evaluations were delivered on the basis of a contingent value survey and a comprehensive calculation of the costs and savings produced by portability policy.
The current policy underlying the portability of Australian pensions reflects international practice. It acknowledges the presumption that payments made in respect of the major contingencies of life—old age, invalidity and widowhood—should continue to be paid regardless of where the recipient lives. The policy also promotes international social security agreements as an effective way of giving former residents access to entitlements that would have been available to them had they remained in Australia.
In 1972, Australia introduced indefinite portability of Australian age, invalid, widow and wife pensions under bilateral portability treaties with Greece, Turkey, Italy and Malta. These treaties arranged the indefinite export of those pensions to the specified countries, subject to strict residence and qualifying event conditions.3
In 1973, in acknowledgment of the aspirations of migrant communities, Australia unilaterally introduced indefinite portability of many of its pensions. Indefinite portability was extended to most pension categories and imposed no additional residence conditions for export. Generally, if a pension was payable in Australia, it was payable overseas. The new approach obviated the need for the earlier portability treaties, although those treaties were not formally terminated until later. The approach was supported by the 'special needs' provisions so that former residents with a substantial connection to Australia who were in special need of financial assistance could access entitlements if they had left Australia without a portable pension before the enabling date of the new legislation (8 May 1973).4
The evolution of portability policy since 1973 has continued to defer to the residence-based nature of the Australian system. In the past decade, policy initiatives have curtailed the export of pensions acquired with very little residence. In the context of an agreement network based on the principle of shared responsibility, policy enhancements, such as proportional portability, have sought to rationalise the extent of Australia's responsibility to provide income support to non-residents. The important changes that have been made to the 1973 general portability provisions are as follows.
- On 8 May 1985, consistent with its intention to create a network of international social security agreements, the Government modified the general portability conditions to introduce (with savings provisions) 'proportional' portability for pensions granted after 1 July 1986.5 Based on Australian working-life residence, proportional portability is a mechanism to control the rate of the portable payment. It was introduced to underpin the creation of a network of shared responsibility agreements with countries that provide contributory entitlements. Australia's actual and prospective agreement partners generally have contributory social security systems that export pro-rata benefits, based on the individual's contribution records. Proportional portability enables Australia to reciprocate on a similar basis, as the amount of pension for export is directly related to the period of residence in Australia during a person's working life.6 The maximum means-tested entitlement is obtained after 25 years of working life. An individual with 15 years' working life residence in Australia could expect to receive a proportional pension rate representing 15/25ths of the means-tested entitlement.
- Indefinite Carer Pension portability was stopped from 1 October 1987, except under some international social security agreements.
- From 1 July 1988, the portability of Sole Parent Pension was limited to the first twelve months of an absence except for special widows.7
- Departure certificates were introduced from 1 February 1989.
- From 1 July 1990, the portability of wife and widow B pensions was limited to the first 12 months of an absence unless recipients are 'entitled persons'.8
- From 12 November 1991, a 12-month limitation on the portability of certain Disability Support Pensions was introduced.
- In 1992, short-term portability of Carer Pension was re-introduced for carers travelling overseas together with the pensioners being cared for and for carers in a respite period.
- From 1 January 1993, Additional Family Payment ceased to be portable. When additional payments for children of pensioners were integrated into the family payments system, portability of those payments was barred unless pensioners were already overseas or able to export the payments under an international social security agreement.
- From 21 March 1994, legislation was implemented requiring Australian pensioners to pursue unclaimed entitlements from agreement countries.
- With effect from 1 January 1995, the penalty clauses in the departure certificate provisions were moderated.
- In 1996, portability of Family Allowance above the minimum was enhanced (for the first 13 weeks of a temporary absence).
- In the 1999-2000 Budget, it was announced that the legislative framework of portability would be simplified from September 2000. This reform introduced a standard short-term portability period of 26 weeks for all payments. Departure certificates were also abolished, as were the cumbersome savings provisions.
The number of pensioners overseas long-term continues to rise steadily. However, expenditure on pensions paid abroad, as a percentage of total outlays, has not increased significantly over the past seven years. The increase in numbers concurrent with a containment of expenditure can be attributed to the effectiveness of the international agreements program, which gives a larger number of former Australian residents access to Australian pensions but ensures that a greater percentage of payments are made on a proportional, shared basis.
Cost-benefit methodology is viewed internationally as an appropriate way of selecting a socially acceptable course of action. For example, the Swedish Government used cost-benefit methodology to determine the provision of public goods (Corne & Sandler 1986). This methodology proved successful because it allowed a good balance to be struck between one group that had incentives to under-report their willingness to accept the change and another group that tended to over-report (ibid. p. 276). Also, for research on the care of the elderly the Essex County Council selected the cost-benefit method because it allowed:
to bridge the often unfathomable chasm between the norms, experience and data which are the everyday working material of the social worker and the norms, experience and data which are everyday working material of the accountant (Wager 1972, p. 5).
In the analysis of portability policy, a balance between two pairs of elements is sought: between the opinions of people who strongly appreciate portability policy and those critical of it, and between the fiscal and attitudinal effects of the policy.
It was precisely because of a proven ability to establish balance that the cost-benefit method was selected for the analysis of portability policy.
Social security payments are often the subject of emotive public discussion. Some groups argue for a reduction of income support payments and refer to the so-called 'dole bludger syndrome'. Conversely, welfare lobby-groups demand increases in the availability of social security payments and their payment rates. These views are the poles of a wide continuum of community views that, in the context of policy change, are difficult to balance. Through comparing the social benefit that the policy produces with the cost of the policy, cost-benefit methodology has a moderating influence.
Under the cost-benefit methodology, the choice between alternative courses of action is made by reference to the net social benefit that they produce (Department of Finance 1997, p. 1). The action that produces a larger social benefit and costs less is usually selected over other actions that require more investment and generate a lower level of public satisfaction.
In this research, the cost-benefit methodology was used not for selecting an appropriate action, but for evaluating the existing portability policy. Such an evaluation required that an assumption be made that portability policy does not exist and that the money spent on portable pensions could be used for other purposes. As discussed earlier, the introduction of portability was a political decision in response to the international and social situations of that era. If such a decision is taken, it is appropriate to ask whether it should be taken, whether the proposed policy is optimal and whether it should involve legislative restrictions (Baldwin 1995, p. 292). This research does not answer all of these questions. It does, however, throw some light on the community's perception of portability policy and whether the existing modus operandi of the policy should continue or whether some changes should be introduced to address community concerns.
In their critical overview of the role of cost-benefit analysis in improving the efficiency of public sector investment appraisal, Giardina & Williams (1993) concluded that cost-benefit analysis may be considered a useful, informative system for public opinion that makes evident the value judgements underlying the political decision-making process. They also argued that the analysis may reveal inadequate bureaucratic operations and, because information on 'social profitability' is available, also reveal the consequences and effectiveness of a policy (pp. 142, 148, 154-157, 168-169).
Portability has proven to be a contentious issue ever since its introduction. Public opinion was polarised on whether to pay pensions to former residents and long-term travellers. Very little was known on whether the general public supported short-term portability. Anecdotal evidence was conflicting and could not sustain any reasonable assessment of portability. Previous research has shown that portability generally produces savings but no data were available on the level of satisfaction or depth of criticism of portability policy by the general public. Cost-benefit methodology was selected over other research methodologies because it allows for a comparison to be made between the costs of the policy and the level of social satisfaction with the policy. In this way, cost-benefit methodology allows a reckoning to be made of the total net benefit of policy.
This approach, of balancing spending with the level of social satisfaction, is particularly relevant in the area of welfare research. In welfare research there has always been a dichotomy between analysing phenomena in terms of an individualist and a structuralist approach (Popay & Williams 1999, p. 157). Cost-benefit analysis does not transcend this dichotomy but allows both approaches to be put in the context of a single measuring scale. Cost-benefit methodology is able to measure more than just the direct financial benefits of a course of action. It is a good tool for evaluating the net social benefit versus the cost of a policy, such as the portability of social security payments.
At the core of the cost-benefit methodology is a comparison of the social benefit of the policy with the costs of the policy, on equal terms.
In cost-benefit methodology, the cost of a policy is the value of what is forgone in order to implement the policy. Often this cost is defined as the 'opportunity cost' and is understood to encompass a broader spectrum than solely the expenditure of money. Costs are true 'opportunity costs' if they can be saved and used for other purposes (Anderson & Williams 1975, p. 55). Also, when one decision-making process imposes costs or confers benefits on another unit, without having to take them into account itself, the situation is said to give rise to externalities. For reasons of complexity and the low relevance to the objective of the project mostly opportunity costs will be considered and only some alternative costs.
The costs of portability policy will be calculated as opportunity costs that could be saved and these funds then used for other purposes. A saving is made only if social security customers would have gone overseas, regardless of portability policy. In this case the real cost of portability is the amount that would be saved because people would go overseas without their pensions. In other words, without the introduction of portability some pensioners would still travel but the expenditure on their pensions, as well as some infrastructure costs could be saved.
To learn more about the social benefit of portability policy, a specially designed survey was required. The survey had to provide three types of information. The first type was demographical information that could be used to identify various cohorts. The second type was attitudinal information that would allow a classification of the various attitudes towards portability. The last and most difficult type was information that could be used to derive an estimate of the social benefit attributable to portability policy and translatable into terms compatible with the costs of the policy.
The survey design will be discussed in chapter 2. However, it should be mentioned that the survey involved a hypothetical market situation. Respondents were asked to declare how much they would be prepared to pay (in the form of an insurance premium) to continue to have their pensions payable while overseas. It was assumed that individuals who were asked to evaluate portability in the prescribed hypothetical situation would economise just as they would for all other private goods in their possession (Nas 1996, p. 41). Costs of the policy were calculated as if portability policy had not been implemented. The dollar amount declared by survey respondents to have their pensions paid overseas, having been made on the assumption that portability policy was not in existence, thus constitutes a compatible figure with the costs of the policy. Also, the inclusion of respondents who are not currently recipients of Age Pension and those unlikely to be recipients reduces the level of 'political petition' through which, for example, a pensioner cohort would try to influence government decisions.
The cost of portability policy was calculated as the total cost. It was calculated as the yearly cost of paying portable pensions to all customers travelling overseas. The dollar value of the total benefit obtained from the survey was then extrapolated to provide the total in terms of the Australian adult population. Detailed reasons for choosing such an approach are discussed in chapter 2.
In the final chapter of this research, the actual cost of portability policy is compared to the amount of money people were prepared to pay for portability policy. If the cost of portability policy were higher than the amount that the general public declared for it, portability would be costly and would not generate the expected level of public satisfaction. Conversely, if respondents declared a higher amount for the policy than the cost of the policy, then the policy is effective and is appreciated by the general public.
Portability policy is not new and its operation depends on processes that are beyond policy control. Globalisation, migration, overseas social security policies and the economic situations of countries around the world impact on the modus operandi of portability policy. Portability policy therefore needs to be readjusted to these changing conditions. Portability policy operates in a context in which changes are difficult to predict. Depending on the nature of such changes, expenditure could rapidly and unexpectedly grow, for example if the number of Australian social security customers leaving the country increased. It could also produce some savings if the cost of support for the aged in Australia increases. In these circumstances, the cost-benefit methodology helps to base the policy on the overall balance of social costs and benefits. It provides a basis for assessing whether the spending brings about the desired social benefit and how to develop the policy to increase its social benefit.
Portability policy is often discussed in the context of immigration and multiculturalism policies. It is often viewed as a policy beneficial to the migrant community. Paying Australian social security pensions to people residing overseas is also often considered both as a burden on the tax system and as unjustified.
On the other hand, previous research argues that when a person moves overseas considerable savings result from reduced spending on infrastructure support such as health care and community services. Therefore,'by returning to their country of birth, pensioners are not doing Australia a disservice, a commonly held belief, but rather are sharing the burden of their age care between Australia and their country of origin' (Couanais 1993, p. 21). However, while portability policy produces some savings, these savings may not be enough to address some social concerns, for example, that if the policy were to be abandoned altogether the savings would be much higher.
This project allows a comparison of savings or expenditures associated with the policy with the total social benefit it produces. The methodology also places the evaluation in the hands of actual or potential consumers of the policy. The general public may distrust expert opinion. In the case of cost-benefit methodology, the evaluation shifts from the expert to the general public.
Cost-benefit methodology also has shortcomings. In relation to portability policy in particular, it requires that some hypothetical assumptions be made, thereby introducing variables that may not have the weight of reality. For the purpose of this research the assumption to be made is no portability of Age Pension. Respondents were asked to consider the value of portability and to indicate how much they would be prepared to pay for portability rights. On the other side of the equation, the costs of portability policy were calculated as savings that could be made if the policy were abandoned. Because the same hypothetical assumption of the non-availability of portability appears on both sides of the equation (costs and benefits), in the final mathematical calculation this assumption can be disregarded.9 On the basis of such a calculation we can create a model that allows the determination of both the value of the social benefit and the net cost or benefit of the policy in a changing environment.
The basic assumptions for evaluating social benefit on the basis of a contingent value survey are:
- Respondents can and do take the initiative to determine their preference orderings between the public goods and services.
- Respondents will not behave strategically, that is declare payments that are intended to bias survey results, rather than reflecting their true preferences.
This evaluation method does have some shortcomings. There is the possibility that, due to survey design, responses could be biased. This bias could flow from the information provided to respondents or from respondents being uncertain of the value of portability policy. Also, given that the situation they are asked to evaluate is hypothetical, some respondents may lack the initiative to perform the introspective analysis required to determine their true preference ordering. Also, iterative bidding may bias results by forcing subjects into stating higher bids. Most of these problems will be discussed in chapter 2. Particular attention will be paid to identification and elimination of structural and hypothetical bias. It should be emphasised, however, that the survey was designed to reduce as far as possible all of these shortcomings. For example, forcing respondents into bidding a higher amount of money was partly eliminated by two sets of questions. Firstly, respondents were asked to declare fortnightly fees of $2, $4, $8 and $12. Then they were asked what was the maximum amount they were willing to pay. In the second set of questions they could revise the previously declared amount and once more they were reminded of their hypothetical financial situation.
There are also additional difficulties posed by the fact that a number of respondents may know that portability policy is a 'free' social service and completely unrelated to the market. Therefore, they may declare completely unrelated values. Indeed, a major criticism of the evaluation of social benefit, based on the contingent value survey, is that individuals are aware that the provision of public goods is not based on a hypothetical statement of willingness to pay. Provision of such services typically result from political process and lobby group pressure. Therefore, exercises designed to elicit hypothetical declared values may not be taken seriously (Cox et al. 1986, p. 170). This possible bias is very important for the understanding and interpretation of the social cost evaluated on the basis of the survey. To create a convincing hypothetical market situation, financial and 'stage of life' situations were described to respondents and restated in crucial contingent value questions. In this way, individual preferences were weighted by an intensity factor that correlated with the individual's income (Pearce & Nash 1981, p. 10).
The strength of the cost-benefit method is that it allows an assessment of social benefit produced by policies that otherwise would be difficult to compare with expenditure and that it produces values that 'usually compare well (are consonant with) analogous values obtained from alternative, market-based methods (Cox et al. 1986, p. 166). In the case of portability, the application of any other market research method would be difficult, if not impossible.
In this research three major steps will be taken:
- On the basis of the survey (a representative cross-section of Australian society), the social benefit of portability policy will be evaluated in monetary terms.
- On the basis of statistical data, the total cost of portability policy including some spill of costs will be calculated.
- The general public's evaluation of portability policy (the social benefit) in monetary terms will be compared with the costs of portability policy.
- 2.1 Contingent value survey
- 2.2 Survey population
- 2.3 Total portability amount
- 2.4 Calibrating the survey data
- 2.5 Interpretation of the total social benefit amount
As previously discussed, cost-benefit methodology requires that all values be expressed in monetary terms. A monetary expression of non-market goods is often difficult and requires complex abstract assumptions. To evaluate the social benefit of overseas portability of Australian social security payments in monetary terms is relatively straightforward in that is does not require complex abstract assumptions.
The monetary value of the social benefit may be derived via the creation of an artificial market situation. Goods not normally subject to the market are defined in market terms. For example, air quality is not a market commodity. However, if a factory were to be built near houses, pollution would likely increase, as would the cost of providing health services. Considering the effects of this pollution, air quality can be expressed in monetary terms and compared to the benefit produced by building of the factory. Generally, to evaluate the cost and benefit of the building the factory, non-economic factors such as life safety, health and environmental quality will have to be assessed. Individuals should be asked how much they will gain and how much they will lose if the factory were constructed. The basic assumption of this reasoning is that:
If we accept the fundamental value proposition that the society does not have an independent existence apart from the individuals constituting it, social welfare should be a function of individual welfares or preferences. Then the cost-benefit analysis can be done by estimating the relevant quantities of different items of gain or loss in terms of their importance in individual preferences and to compare these individual preferences in terms of the social welfare function (Ng 1993, p. 5)
To investigate these non-economic factors, respondents of a specially designed survey are asked to consider how much they would be prepared to pay for services (the 'willingness to pay' technique) or how much they would want to be compensated were the services to be withdrawn (the 'willingness to accept' technique).
Portability of pensions may become a market good if we consider that to have the pension paid while overseas a pensioner has to insure it, in a similar way as people take insurance against loss of income. In a specially designed contingent value survey two assumptions, which will be discussed below, were taken:
- the 'willingness to pay' technique will provide better results than the 'willingness to accept' technique; and
- the survey population should be a randomly selected representation of the Australian population.
To evaluate the total social benefit of portability policy, the 'willingness to pay' technique was selected over the 'willingness to accept' technique. This was primarily because it was assumed that respondents would find it easier to think about the value of having a portability option, rather than to think about compensation were portability not available. Also, the 'willingness to accept' technique would create a situation in which most, if not all respondents would have reason to demand very high compensation. The three possible groups of respondents would each have an incentive to demand high compensation. Respondents not wanting to travel and those not intending to travel in the foreseeable future would demand high compensation because it would increase their overall income without compromising anything of benefit to them. Also, respondents intending to travel would demand high compensation. There would not be any real differentiation between these groups. Each group would have a good reason to demand the highest possible compensation.
Generally speaking, travelling constitutes an important part of contemporary life. However, it is not a life function that results in substantial personal trauma when withdrawn from someone. To test the 'willingness to accept' technique the survey would have to use a hypothetical situation that a customer has to go overseas, for example to attend to an acute family crisis, but the pension is not payable. Such a situation would reduce the level of neutrality and force respondents to accept payment for portability. In contrast, application of the 'willingness to pay' technique does not create a situation of constraint and allows for a free choice of options.
As indicated in chapter 1, portability policy allows social security customers to continue to receive an Age Pension while overseas. To evaluate this policy two approaches can be taken: an evaluation of portability policy in relation to age pensioners, or an evaluation of portability policy in relation to the general public.
The first approach was rejected for a number of reasons. Age pensioners are the immediate beneficiaries of the policy; as such their evaluation could be biased. Also, we found that age pensioners did not unanimously appreciate portability policy. Age pensioners may not be interested in travelling, or may even be resentful of those who can or want to travel. They may harbour strong feelings about immigration policy or other political issues. By targeting just age pensioners we would only derive an evaluation from people who are already the beneficiaries of portability policy and the opinions of other groups of people would be missing. It is also important to minimise the possible structural bias that comes from a particular social position or interest of respondents (Nas 1996, p. 111).
An evaluation derived from a survey of a randomly selected sample of the Australian public is comprehensive and also provides a sounder basis for future development of the policy. Since the evaluation includes the opinions of people both directly and potentially affected by the policy, as well as those people unaffected and possibly critical of it, the scope of opinions allows for a comparison and a more comprehensive interpretation of the contingent value questions. This is particularly important because portability policy affects not just age pensioners but also recipients of other payments, including family payments, Sickness Allowance or Newstart Allowance. It is the right of any Australian resident who is in genuine need to receive an income support payment. This right is not linked to contributions to a social security fund or the payment of taxes. Everyone can potentially be, for a period of time, a recipient of a social security payment. In this context, only the indiscriminate treatment of recipients, potential recipients and non-recipients of any age and any social strata will allow for a comprehensive assessment of the social value produced by portability policy. Were any of the above groups excluded, the benefit of portability policy would be narrowed to this particular group, representing a specific and often temporary interest in portability policy. By considering all groups it is possible to estimate the social value of portability policy from the viewpoints of various groups and interests.
However, responses from the general public may not be free of bias. Opinions held by the general public may be distorted by hypothetical biases arising from unconsidered evaluations or from strategic biases related to unimaginative evaluations. For various reasons, often relating to idiosyncratic characteristics, answers to contingent value questions may differ from a respondent's true value for the goods. Dealing with this problem usually requires complex statistical operations that, in most cases, do not produce straightforward results (Mansfield 1998, p. 665). Possible hypothetical and strategic biases are discussed in the following paragraphs. Meanwhile, it is worth noting that, although hypothetical bias may be higher for the general population than for the age pensioner population, its effect could also be counterbalanced by a lower strategic bias caused by the specific situation of age pensioners. In any event, the comparison of various cohort and demographic data with specific evaluations helps assess some of the bias.
The primary aim of the survey was to answer the question:'How much does the general public value the portability of the social security age pension?'. Interviewers made sure that respondents understood the contingent market scenario and encouraged and reminded respondents to carefully follow the scenario that had been described to them. The survey questions were designed to meet the methodological imperative that 'the scenario has to be understandable and meaningful to the respondents and free of incentives which might bias results (Carson & Mitchell 1989, pp. 17, 297).'
The questions were structured to identify both 'yes/no' positions and the intensity of the preferences held by respondents. Intensity was measured by asking how strongly respondents agreed or disagreed with a particular opinion, or how important a particular option was to them.
A fundamental objective was to collect the opinions on portability of age pensions. Most other social security payments are portable. However, age pensions can stand proxy for portability of other payments. Unlike other payments Age Pension portability is unrestricted. It is not only much easier to speculate on the length, reasons and destination of travel but also to see it in the context of retirement. Most people, including young people, occasionally think about what they are going to do in retirement. It could therefore be assumed that the concept of Age Pension and its portability in the context of retirement is clear to the general public. Also, that a contingent valuation survey which refers to this well established public knowledge will not be vulnerable to a temporal selection bias, that is that respondents would be too concerned with their existing situation and too strongly influenced by their temporal interests.10
People living overseas were not interviewed. It is likely that age pensioners who are living overseas and receiving their pensions would value portability policy more highly than any other group. Because this group derives a direct benefit from the policy, inclusion of this group would shift the balance and unduly inflate the results.
The project team developed a questionnaire for portability policy consisting of three parts: one dealing with demographic characteristics, one with attitudes toward portability policy, and a portability evaluation. The evaluation was based on individual behaviour in a contingent market setting. Every respondent was introduced to the same hypothetical market situation and was asked the same set of question regardless of any demographic or attitudinal characteristics identified in the early questions.
Through an open tender process the team sought a consultancy to conduct the survey and interpret the survey data. Bidders were provided with an outline of the questionnaire and the research method. After a thorough selection process, Roy Morgan Research was awarded the contract to conduct the survey and calculate the total social benefit.
Before the survey, three pilots were conducted to ensure that the questionnaire satisfied the research objectives and that the questions were easy to follow. The latter requirement was particularly important because of the complexity of the contingent value questions and their political sensitivity. As a result of the pilots, it was decided that the questionnaire and the contingent value questions should be introduced after a brief explanation of their purpose. The results of the pilots do not form part of the final report.
During March and April 1999 Roy Morgan Research interviewed 1 009 respondents. The sample size was calculated on the basis of the following formula:
n = 1.962 p(100-p) / SE2
n = sample size
SE = tolerable standard error, expressed as a whole percentage.
The tolerable margin of standard error was assumed to be a rigorous 3 per cent.
p = characteristic being measured, expressed as a whole number derived from the percentage of yes answers to a yes/no question, in this case the question:
Do you wish to travel? 75 per cent yes (confirmed from the pilots).
n = 1.96 x 1.96 x 75 x 25 / 9 = 800
To further ensure robust results the survey sample was increased to 1009 randomly selected households with one person providing responses to the survey.
On the basis of the sample and later, the survey results, the standard error was calculated by using the following formula:
p = percentage survey estimate, as above
n = sample size
The standard error is 3.1 points (at 95 in 100 confidence level), meaning that the results of the survey have to be interpreted within an error range of +/- 3 per cent.
Households were randomly selected from the latest version of the electronic White Pages and only one person from the household was interviewed. Measures were taken to ensure that the sample included an appropriate cross-section of ages, professions, incomes and ethnic groups, as well as an adequate representation of people from rural and metropolitan areas. The final data were weighted to represent the Australian population, using one of Roy Morgan Research's standard weighting schemes that uses age, sex and area, and requires an 88 cell matrix (11 x 4 x 2) consisting of 11 areas by 4 age groups by sex.11 In the report the resultant weighted cases and the number of responded represented by each percentage are discussed. It was considered that, so long as the sample is weighted to reflect the demographic composition of the Australian society, the results of the survey could be extrapolated. The size of the sample was validated with the statisticians at Roy Morgan Research and in consultation with Department of Sociology of the Research School of Social Sciences at the Australian National University. Also, a majority of researchers consider that an increase in the size of a sample is not of itself sufficient to guarantee increased accuracy of results (Kalton & Moser 1933, pp. 146-47). On the contrary, an increase in sample size decreases the error by so little that it is not usually worth the additional costs (Singleton et al. 1993, p. 169).
To achieve more meaningful responses we skewed the sample by setting quotas to ensure that a larger proportion of respondents were aged 50 years and over. This procedure has caused some loss in the overall precision of the sample. However, the loss of precision, around 7 per cent, allowed us to increase the precision of the estimates in the contingent value question by around 20 per cent. Overall, as can be seen in the following table, the sample population reflected quite well the demographic characteristics of the Australian population.
|Age||ABS data||Survey||Standardised survey data|
The questionnaire was programmed and respondents interviewed using Computer Assisted Telephone Interviewing (CATI). Up to four attempts were made to contact respondents. After an unsuccessful fourth call, another respondent was randomly selected. All calls were made between 6:00pm and 10:00pm. It was expected that contacting young people during the evening might be difficult. Therefore, the youngest person in the household (over 18 years) was asked to respond at a mutually agreed time.Where possible, an interview time was arranged.
The contingent value survey has proven to be a very useful method in gathering data for an evaluation of the social benefit of portability policy in monetary terms.
To evaluate the total social benefit, five 'willingness to pay' type of contingent value questions were asked. Respondents were informed about the hypothetical character of the questions and the conditions of the imagined context. They were also assured that the questions were being asked for research purposes only and did not reflect any policy intentions on the part of the Department of Family and Community Services.
In the first contingent value question, respondents were asked to imagine a situation where they were retired with an Age Pension of $400 per fortnight as their only income. In order for their pension to be portable, that is, paid while they are overseas, they have to pay a fee of $2 each fortnight. The questions were designed to encourage respondents to view the fee like an insurance premium. It should reflect the value the respondents place on the portability of their pensions. The rate of insurance usually does not depend on the length of contributions. It is usually related to the value of the insured goods or function. However, it is important that one is insured at the time when the event against which insurance has been taken occurs. In this case, the event is characterised as 'going overseas' and when questioned, interviewers explained that the person has to be insured for around six months before departure to have their pension portable. Otherwise the pension is payable only in Australia.When questioned, interviewers explained that this amount must be paid every fortnight for at least six months before departure.
To establish the highest amount that respondents would be prepared to pay, the initial $2 amount was gradually increased. In the final question respondents were asked about the maximum amount they were willing and able to afford to pay. This last question provided respondents with the opportunity to revise their previous decisions, including adopting a decision not to pay.
We will refer to the amounts derived on the basis of the first set of questions that asked respondents to declare a 'set-fee amount' of between $2 and $12. The amount derived on the basis of the last, free choice question will be referred to as the 'revised amount'.
The total social benefit value was calculated by aggregating values declared on the basis of either:
- the highest value declared in the 'set fee' questions; or
- the value declared as the 'revised amount' if that value was higher than the set-fee amount.
Totalling all dollar values from this last question produces the social benefit of portability policy (an amount of $12 572.18). The total social benefit for the whole Australian population was estimated by multiplying this sum by the Australian population 18 years of age and above, and dividing the result by the population of the survey. The total portability amount derived in this way was $175 438 576.
Although the survey data were weighted to reflect the demographic and social composition of the Australian population, this amount only partly reflects the real value of portability policy. As was rightly pointed out in the Roy Morgan report, it may be inflated for the following reasons:
- Half of the total amount was generated by those who did not anticipate receiving an Age Pension.
- Those who did not intend to travel overseas were prepared to pay as much as those who did.
- Around 2 per cent of those who stated that they would not travel overseas put the value of portability at more than $50 per fortnight.
- The estimate of the likelihood of permanent departure may be an overestimate because this likelihood decreases with age.
- Prior knowledge of portability policy or Australian social security agreements has not been taken into account.
- The decision to make a payment may be affected by the remoteness of the event (Roy Morgan Research 1999, pp. 14-15).
Most of these concerns expressed by the Roy Morgan research team will be taken into account in interpreting the total amount. Moreover, careful analysis is required because of the multiplying effect when the survey results are extrapolated to the whole Australian population.
These concerns are discussed in the following paragraphs. In relation to points 5 and 6, it is worth emphasising that the evaluation is based on a hypothetical situation. Respondents were asked to imagine themselves in the situation of an age pensioner on a limited budget. They were asked to evaluate the portability of pensions as if they were directly affected by the policy. For some respondents, this scenario involves contemplating a completely abstract situation.
For other respondents likely to become age pensioners themselves in the future, or for others with a family member in receipt of an Age Pension, the idea of evaluating portability policy as if they were affected by it is a more realistic proposition.
Contingent value surveys used to estimate the value of social benefit in monetary terms are also prone to corruption from biases in the responses. For various reasons respondents often do not provide their true valuation. We tried to eliminate or reduce the impact of these biases on the overall result by determining the influence of demographic and attitudinal characteristics on valuations of portability policy.
To calibrate the data, possible biases must be analysed. The relative amount declared by some respondents may not reflect a realistic value. Answers may be influenced by personal characteristics, political attitudes or on account of respondents' remoteness from the hypothetical scenario they are being asked to consider. The value of portability policy may be overestimated or underestimated because it is not of concern to them or because they have other countervailing views.
When a person thinks about a situation without a 'reality check', hypothetical bias may occur. To discuss this bias, we examined the evaluations of young people, respondents on higher incomes, and those who are unlikely ever to receive an Age Pension.
Conversely, some respondents may not be capable of detaching themselves from their day-to-day situation. Their anchor in the status quo does not allow them to imagine another situation and their evaluations will thus be too realistic. This kind of bias, when a person finds it difficult to abstract, will be called the structural bias. To discuss this bias, we examined the evaluations of respondents unlikely to travel, the elderly, those on low incomes and others who may have been influenced by this bias.
The following factors may have been the cause of hypothetical or structural biases in this study, the purpose of which was to find a realistic estimate of the total social benefit amount.
Travelling overseas permanently or for a long period of time
Respondents thinking about going overseas could provide a portability valuation that substantially differs from the valuations provided by those who are not contemplating travelling. Either hypothetical or structural bias could occur in the valuation of non-travellers. Also, respondents thinking about living overseas permanently may strongly over-value the portability of their pensions.
It is often assumed that a world in which money and information are increasingly mobile is also a world in which people move and live in foreign countries. This assumption has not been proven correct. The survey showed that many respondents intend to travel overseas but only a small percentage intended to travel for a long period of time or to live overseas permanently.
Many respondents reported that they were likely to travel for periods of less than 6 months. Six in ten respondents (59.5 per cent) stated that they would go overseas for periods of between 3 weeks and 6 months. Just over half of the survey respondents (51.1 per cent) would go for periods of between 4 weeks and 3 months.
Respondents who intended to travel for less than 6 months paid the highest amount of money for short-term portability and the lowest for long-term or permanent portability. In contrast, respondents who did not expect to travel when they retire were prepared to pay the highest amount for long-term or permanent portability. Figure 1 shows the significant drop in the amount declared by respondents who intend to travel. This drop reflects a strong appreciation of and support for short-term portability and much less support for longer or permanent portability. In contrast, for respondents who do not intend to travel there was far less differentiation between declared amounts in the context of the length of travel. This may not indicate some hypothetical bias, that is, thinking about the situation without a 'reality check'. In contrast, those respondents may think about travelling without considering the length of travel. It is likely that they evaluate portability by thinking: 'if I travel would I like to have my pension paid and how much should I declare to secure the payment'. It is also important to note that because of the nearly uniform values declared for all lengths of travel the overall amounts produced by this cohort is much higher than the overall amount produced by the traveller cohort.
Figure 1: Total payments by respondents likely and unlikely to travel
Figure 1 also shows a significant drop between the total amount declared by respondents intending to travel for less than 6 months and respondents wanting to travel for longer periods, or to live overseas permanently. Within the group of respondents expecting to travel for less than 6 months, the amounts nominated by respondents expecting to travel for periods of between 4 weeks and 3 months was much higher than the amounts declared by respondents intending to travel for periods of between 1 and 3 weeks. This finding shows that short-term portability, for a majority of people, means periods of between 4 weeks and 3 months. Very few respondents thought about travelling for less than 3 weeks or for more than 6 months.
Figure 2 shows that, contrary to what may be expected to be the case, respondents likely to go overseas for long periods or permanently were not prepared to contribute a large amount of money. All 40 respondents declaring their likelihood of living overseas permanently were prepared to pay no more than $20 per fortnight. In contrast, some among those intending to go overseas for long periods of time but not permanently were prepared to pay more than $50 per fortnight. Also, the percentage of respondents ready to pay a relatively small fee was greater among those likely to go overseas for a long period or permanently than among any other group.
Figure 2: Payments according to length of expected travel
Respondents likely to travel overseas for long periods were prepared to contribute 35 per cent more than those who were not. The drop in the percentage of respondents wanting to go for longer than 6 months, in comparison to those wanting to go for less than 6 months, is equal across the range of demographic characteristics, at around 75 per cent. However, respondents intending to go permanently were prepared to pay an amount that was, on average, 230 per cent less than respondents not intending to go overseas permanently.
The high value of the total payments by those who did not intend to travel indicates that they greatly value the possibility of having the Age Pension paid while overseas. Another possible interpretation is that the high amount declared by non-travellers reflects their view that those who travel should pay more for portability. This interpretation has to be rejected because the survey respondents were specifically asked to imagine themselves in the situation of a pensioner who wants to travel. It is difficult to imagine the non-travellers being so wily as to surreptitiously convey the message of what they think other people should do.
The marked drop in value among respondents wanting to travel and the more consistent valuation of those unlikely to travel demonstrate different thinking patterns. Those who want to travel think in terms of a more precise time scale and are prepared to declare higher amounts for the periods that they want to travel. This is clear evidence of structural bias. On the other hand, respondents who are not thinking about travelling at present declare similar amounts of money, regardless of the length of travel. They probably think about portability in more general or even imaginary terms. Their high estimation may result from the existence of hypothetical bias. Nevertheless, occurrence of both biases does not have any substantial effect on the total social benefit amount. Not only do they neutralise each other, overestimation by non-travellers being balanced by underestimation among travellers, but also both biases may be rationally explained. Together they indicate a high appreciation of the portability of Age Pension.
Immigrants and respondents with family overseas
Other biased valuations could come from immigrants and respondents with family overseas. Immigrants may over-value portability policy because of a longing for their family and country of origin and because they are generally more likely to travel. Also, they may undervalue portability because of some misunderstanding about the non-contributory nature of the Australian social security system and, based on experience in the country of origin, they may consider portability to be an 'earned right'. Respondents with family overseas may be thinking about visiting loved ones and be inclined to overvalue portability. They may also be thinking about the financial support that they would feel obliged to give, or could receive during their overseas visit and undervalue portability. Amounts declared by both of these groups may be strongly polarised.
From another point of view, the migrant cohort could constitute an important factor influencing the valuations made by other respondents. Strong views on and sentiment towards migrants could produce a biased valuation. It could happen among the respondents who associate portability policy with immigration policy.
The survey data show that people born overseas or having family overseas are more likely than any other group to go overseas for periods of longer than 6 months or to decide to live overseas permanently (see Figure 3).
Figure 3: Background of respondents intending to travel
Respondents with families overseas were more likely to expect to travel than any other group. They volunteered a slightly lesser average amount than those who did not have family overseas. Also, respondents who stated the reason for going overseas 'to be close to relatives and friends' were prepared to pay, on average, twice as much as other groups. This could be the result of structural bias: they were thinking about support from or to be provided to their families while overseas.
People with families overseas, like the overseas-born, tended to decrease their initial amount (declared on the basis of set-fee questions) by around three-quarters, whereas other groups tended to increase the declared amount by around half. The average amount volunteered by respondents without families overseas was only slightly higher, by around one-quarter.
The 'family overseas' factor influences the evaluation of portability policy. However, this influence does not make a strong impact on overall declared amounts because overvaluations are often neutralised by undervaluations. Another important factor is that immigrants and respondents with family overseas are more likely to pay when the fee is set than when they have to decide about the declared amount. That some overseas-born respondents paid the set fee of $2, $4, $8 or $12, rather than taking the opportunity to provide a free evaluation of portability, could be explained by previous experience of this cohort. Most overseas social security systems are based on contributions, that is, where the person pays periodically into an insurance type fund. The Australian system is based on residence and need; there is no requirement to make contributions. Therefore, Australian-born respondents evaluate portability by finding the dollar amount that reflects the value of portability policy, whereas overseas-born respondents are more inclined to accept a set fee, similar to a contribution rate. Also, they probably think about travelling in the context of strong emotions. Perhaps, on reflection, they reduce the initial amount in order to increase their disposable income while overseas.
Apart from immigrants and respondents with families overseas, single young male respondents were likely to go for long periods or to live overseas permanently. For example, single respondents were prepared to pay on average almost two-thirds more (62 per cent) than respondents who were partnered or planning to marry. The same pattern occurred in other categories so that, overall, it was not possible to identify a single biasing factor relating to immigrants or respondents with family overseas.
Contrary to the expectation that immigrants would value portability of Age Pension more than the Australian-born, the survey showed the opposite. Almost nine out of ten (87 per cent) of Australian-born respondents increased the set fee of $2. By comparison, only one in eight (13 per cent) of overseas-born increased the initial amount of $2. Also, the average payment of the overseas-born was 40 per cent lower than for respondents born in Australia. On reflection, the Australian-born valued portability policy more highly. Moreover, they were much less affected by emotions associated with national identity or family. Rather, they probably associated travelling with pleasure, interest or curiosity.
As shown in Figure 4, the value declared for short, up to 6 months' portability, does not differ much between immigrant and Australian-born cohorts. The difference increases for longer or permanent portability periods.
Figure 4: Support for portability policy
Four out of five (81 per cent) respondents agreed that Age Pension should continue to be paid to short-term travellers. This strong support was reflected among both immigrants and the Australian-born. Despite similar levels of support for short-term portability, the Australian-born and immigrants strongly differed in their opinions about long-term or permanent portability. Among the Australian-born, over 86 per cent disagreed and only 18 per cent agreed that pensions should be paid to overseas residents. In contrast, among the overseas-born the division was less distinct, with 51 per cent disagreeing and 42 per cent agreeing.
In relation to short absences, Australian and overseas-born respondents expressed an equal level of support. The difference in the support for long-term and permanent portability can be attributed to the stronger interest immigrants have in going overseas for longer periods of time. Since for most respondents travelling meant a short-term overseas stay, the expression 'live permanently' could be interpreted as severing all ties with Australia. This understanding could also differ among migrants, who could interpret this as returning to their country of origin for retirement and continuing to maintain links to two countries. In view of this possible difference in understanding, perhaps residents of Australia have a very strong feeling that social security payments should continue to be paid to people who are travelling but not to people who are living overseas.
Reasons for travelling
Reasons for travelling could have some impact on the evaluation of portability policy. Even more importantly, an analysis of the contributed amount of money, in the context of 'reason to travel' category, could throw some light on whether portability plays an ancillary or a pivotal role in the decision to travel. This role is important because, if people travel for longer periods of times regardless of the portability of their pensions, the total cost of portability policy would increase.
The survey results show that three times as many respondents gave, as a reason for traveling for short periods of time,'to see the world and to experience other cultures', rather than 'to visit families or friends'. However, as expected, for older respondents and the overseas-born, the latter reason is slightly higher than for others, on average around 9 per cent.
A similar proportion exists for people wanting to go for longer periods of time. But 'seeing family and friends' prevails among people wanting to go overseas permanently. It is interesting that large discrepancies between 'seeing family' and 'seeing the world', characteristic for shorter and longer periods of travel, are not repeated for permanent departures. Around 39 per cent of respondents intending to go permanently expressed the desire to be close to family. Similarly, 33 per cent wanted to experience other parts of the world.
Figure 5: Major reasons for traveling
It is worth noting that people who could not travel for health reasons were prepared, on average, to pay 1.8 times more than those who did not want to travel through lack of interest. Clearly, the reason for travelling is strongly reflected in the nominated amounts. Generally, Australians travel because they want to experience the world, whereas the family factor is a strong influence on those thinking about permanent departures.
The declaration by more than 58 per cent of respondents thinking about going overseas for a long period that they would not go if the pension stopped is a substantial indication of the social benefit of portability policy. Within the group of people wanting to live overseas permanently, respondents giving the reason 'to be close to family and friends' are determined that they would not go without a portable pension, as are respondents wanting to experience other cultures and see the world. It is also interesting that respondents over 65 years of age were more definite than any other group about whether they would or would not go without a pension. More than 40 per cent of them would go, but around 60 per cent would not go.
The financial situation of respondents is probably the most important differentiating factor of all. It is reasonable to assume that people on higher incomes would have a different perception of money matters than people on lower incomes. People on high incomes may be prepared to contribute amounts that do not necessarily reflect a realistic value for portability. The survey has only partly confirmed this assumption.
Contrary to expectations, half as many people on incomes below $10 000 a year were not prepared to pay for portability (9.5 per cent) compared to people on higher incomes (19.1 per cent). More than 70 per cent of people on low incomes stated that they would go overseas even if their pensions were not portable.
Respondents not in paid employment came fourth on the list of the highest contributors, after professionals, the skilled and those in sales work. There were cases of people on low incomes with no skills where the declared amounts did not seem realistic, for example, more than $140 per fortnight. By comparison, the next highest payment was declared by the professionals ($28.70 per fortnight). Therefore, to limit the potential for these seemingly unrealistic amounts to skew the results, the implied amount of $3 571 000 for people on low incomes with no skills will be disregarded in the calculation of the total portability amount.
On average 37 per cent of the respondents not expecting to travel gave financial or health reasons for their decision. Among this group, immigrants more than any other group gave financial difficulties as the reason for not travelling (more than 55 per cent).
A comparison of 'likelihood to live abroad permanently' with approximate annual income showed that respondents with incomes of between $20 000 and $24 999 were the most likely to go overseas permanently. The overseas-born constitute only 8.9 per cent of this income group rather than their average 21.3 per cent.12 This partly explains why migrants contributed less than the Australian-born. According to the survey data, migrants belong either to a low-income or high-income group, so they are influenced in their decision making by their financial situations.
The elderly, and respondents receiving or likely to receive Age Pension
Imagining themselves in the situation of pensioners wanting to travel overseas is easier for some respondents than for others. Over half said they were likely to travel, including around one in five respondents identified as already retired or semi-retired. However, for more than three-quarters of the survey population, the thought of retiring and travelling is still a matter just for speculation.
Another important point to consider is the fact that respondents were asked to assume the non-existence of portability policy. Some respondents, familiar with the portability of Age Pension, may have considered the requirement to pay foreshadowed a change in government policy. It is possible that these respondents were willing to declare only a very small amount of money. They may have feared that the amount they declared would be taken as an indicator of what fee people might be prepared to pay for portability policy. Others still may have been thinking purely in terms of securing payment overseas, without due consideration of the limited financial means of the average pensioner. Thinking that even the smallest amount received while overseas to be better than none, they may have declared quite unrealistic fees.
As expected, the survey data confirmed that younger people are more likely to travel than older people. Figure 6 shows that younger respondents (those between 18 and 34 years) were nearly 6 times more likely to travel for a short period than respondents over 50 years of age. However, 'intention to travel' increased again among respondents over the age of 59 years. This trend may reflect changing lifestyles, with enthusiasm for travelling declining with age. Enthusiasm for travel increases again when people think about retirement but wanes when failing health makes travelling difficult.
It is interesting that no respondent aged between 50 and 64 years was thinking about going overseas for a period longer than 6 months. Few respondents younger than 49 and older than 65 years were prepared to go overseas permanently.
Figure 6: Comparison of age and number of people willing to travel for relevant periods of time
As expected, respondents between 50 and 54 years of age, a group that would be thinking seriously about retirement, were also prepared to pay the highest average amount for portability. In second position were people between 35 and 39 years of age. People between 65 and 69 years of age ranked third.
This trend is not repeated among respondents likely or unlikely to receive age pensions. There is no clear difference in the pattern of payments nominated by respondents receiving Age Pension and those who are both likely and unlikely to receive them. Only a small difference (8 per cent) existed in the average payments nominated by respondents expecting to receive Age Pension and by respondents who did not have that expectation. By comparison, the average amount paid by social security customers was around one-fifth lower than others. The average amount derived from the open choice question did not differ among recipients and non- recipients of social security benefits. However, semi-retirees were prepared to pay more for portability than retirees and those in employment.
Respondents thinking about retirement and not necessarily about getting an Age Pension value portability more than any other group.
Figure 7: Likelihood of getting an Age Pension and amounts people declared willing to make
These analyses demonstrate the existence of structural and hypothetical biases in the evaluation of portability policy. At the same time, it is difficult to distinguish any group characteristic that would warrant a significant adjustment to the total amount. Most respondents understood the contingent value questions and gave plausible evaluations. Also, as the analyses have shown, even where the evaluation appeared over or understated, reasonable explanations could be provided.
Roy Morgan Research calculated the total social benefit amount. Their technique of calculating was based on a simple methodology. The total sum of maximal payments across the whole sample was derived on the basis of the highest payment declared by respondents when they were asked whether they are prepared to pay $2, $4 and up to $12. Also, if the value declared in the free evaluation question was higher than in set fee questions this value was taken into account. The total amounts were weighted and the total sum of weights was 1 009, with the average weight being equal to 1.0. The total sum derived in this way was $12 572.18.
To derive the total sum in respect of the total Australian population, the population figure of 14 080 100 was used (18+ yrs, ABS December 1998). Thus, a population estimate was divided by the size of the sample to produce a factor, 13 954.51 (14 080 100 + 1 009 = 13 954.51). Thus, the total portability amount for the Australian population calculated by Roy Morgan is equal to the product of the factor and the total sum derived from the sample and equals $175 438 576. However, as discussed earlier, $3 571 000—the amount declared by the 'no occupation' cohort—is subtracted from the total amount calculated by Roy Morgan. The final Roy Morgan figure then is $171 867 576.
It has to be emphasised that this amount was delivered on the basis of the survey that created a hypothetical situation in which respondents were asked to evaluate the portability of Age Pension. This value does not have a real meaning and will only be used for a comparison of the costs and benefit of portability policy on equal terms.
An important facet of the evaluation of portability warranting further discussion was the respondents' declaration of amounts higher than the required $2 fee. In a normal market situation once goods can be purchased for a lower price it is unlikely that people will seek to pay a higher amount. However, it is also a well known phenomenon that people are willing to pay more if the purchase of the goods fulfils another purpose, for example, the purchase of something where the funds will benefit a school, hospital or aid agency. In such cases the price of the goods does not just reflect their market value but also includes an additional or contextual value attached to those goods. Respondents who could 'purchase' portability for the $2 fee and who were prepared to increase that fee showed that for them portability has an extra value. This extra value could include the possibility to travel the world, to visit family, or to learn about other cultures. For this reason we will focus on the rate of increase of the initial $2 amount.
The relative number of people willing to contribute more than the required amount of money could indicate the level of appreciation. A large number of respondents were willing to pay higher than the initially required amount of $2 per fortnight. Among these, young people between 18-24 years of age and respondents between 35-39 years of age were nearly twice as likely to increase their maximum amount. Also, people on lower incomes (less than $30 000 per year) tended to increase rather than decrease their initial amount.
Figure 8: Superimposition of the total amounts and percentage of contributing respondents
For all demographic groups, the steepest increase in value occurred in the bracket of payments between $3 and $5. The strongest decline was for payments of more than $50. However, it should be noted that smaller increases in payments require a large number of contributors to produce high total amounts. In contrast, a smaller number of respondents prepared to pay high fees produces high total amounts. Figure 9 demonstrates this situation.
Figure 9: Comparison of the amounts respondents are prepared to pay with the percentage of total respondents
The chart shows that the 17 per cent of respondents who declared $8 also contributed the highest amount to the total amount declared by all respondents. In other words, the drop of 44 per cent in the number of respondents (columns 1 & 2) prepared to pay $8 translates to a 58 per cent increase in the total amount. This also means that 72 per cent of respondents declared an amount that was 27 per cent lower than the amount declared by the highest fee-paying respondents (28 per cent of the total number of respondents). This analysis demonstrates that if we view a fee increase as indicative of a strong valuation, then the total sum should be considered deflated rather than inflated.
The above discussion suggests that five amounts should be considered in the calculation of the total social benefit, as per table 2.
|1.||Roy Morgan amount||(AP/SP)* x SA**||$175 438 611.53||This is a basic amount|
|2.||Amount paid by the respondents prepared to contribute $2 fee on the basis of the fixed-value questions||(AP/SP)* x SA** x 0.765 x $2||$21 517 854.42||Amount which reflects the threshold value of portability policy|
|3.||$2 amount paid on the basis of open-value question||(AP/SP)* x SA** x 0.158 x $2||$4 449 311.79||Amount which reflects the minimal value of portability|
|4.||Total amount contributed on the basis of the open-value question||See Table 3||$122 102 632.32||Amount declared after revision|
|5.||Total amount contributed on the basis of the fixed-value question||See Table 4||$68 988 080.22||Initial amount|
* Australian population factor AP/SP = 13 954.51
** Morgan's sample amount SA = $12 572
|Amounts declared||Amounts taken for calculations||Declaring||Number of survey respondents||Amounts in the survey||Amounts in the population|
|Up to $2||$2||15.80%||159||$318.84||$ 4 449 311.79|
|$3-$5||$4||24.80%||250||$1 000.93||$ 13 967 459.79|
|$6-$10||$8||16.30%||164||$1 315.74||$ 18 360 451.17|
|$11-$50||$31||15.60%||157||$4 879.52||$ 68 091 366.45|
|Over $50||$51||2.40%||24||$1 235.02||$ 17 234 043.12|
|Total||99.90%||1008||$8 750.05||$122 102 632.32|
|Number of respondents||Survey amount||Total population amount|
|Pay $2||771||$1 542.00||$21 517 854.42|
|Pay $4||622||$2 488.00||$34 718 820.88|
|Pay $8||335||$2 680.00||$37 398 086.80|
|Pay $12||203||$2 436.00||$33 993 186.36|
|Total amount||$9 146.00||$127 627 948.46|
|Calibrated total amount*||$4 938||$68 988 080.22|
* The total amount was calibrated by eliminating payments by respondents who did not progress to the next payment. See the text below.
Two major amounts should be considered in the discussion of the total social benefit of portability policy. Below we discuss the meaning of both amounts and their possible use for calculating the total cost-benefit balance.
The first contingent value question asked respondents whether they were prepared to pay a fortnightly $2 fee for the portability of their pensions. The questions that followed increased this amount to $4, $8 and $12.
The population of respondents paying the $2 fee constitutes a group that is progressively reduced as the fee increases. The simplest way of calculating the total payments would be to take the number of respondents prepared to pay a particular fee and multiply it by the fee. However, this calculation would not be accurate because respondents prepared to pay higher fees were also prepared to pay lower fees. The higher fee payers are also included amongst the lower payers. For example, respondents who were prepared to pay $12 are also calculated as payers of $8, $4 and $2 fees. In the total they are calculated as if they were paying $26.
Table 4 presents the total social benefit calculated on the basis of the increasing value questions. The final results were adjusted to exclude multiple payments.
The second amount is delivered on the basis of the maximum value question. This amount does not have repetitions similar to the amount calculated on the basis of the fixed-fee questions. Roy Morgan calculated this amount by taking the maximum payments from the set fee question and, where necessary, the maximum value from the following revised value questions. Another way of calculating this amount is by taking the average payment within the particular group of payments and multiplying it by the number of respondents who declared it. Table 3 shows the results of these calculations.
If we compare the Roy Morgan amount and the amount calculated in table 3, it appears that the former amount is 44 per cent higher. Further, when we compare the Roy Morgan amount with the amount calculated by taking maximum values within each group in table 5 it is $3 520 548 higher. After subtraction of 'unreasonable payments' discussed earlier, the total amount is only slightly higher (by $50 452).
|Amounts declared||Number of survey respondents||Value||Amounts in the survey||Amounts in the population|
|Up to $2||159||$2||$319||$4 449 312|
|$3-$5||250||$5||$1 251||$17 459 325|
|$6-$10||164||$10||$1 645||$22 950 564|
|$11-$50||157||$50||$7 870||$109 824 785|
|Over $50||24||$51||$1 235||$17 234 043|
|Total||1008||$12 320||$171 918 028|
The most interesting aspect is the rate of increase from the threshold amount of $2, which secures portability, to the additional amounts people were prepared to pay. This increase provides the additional figure for the evaluation of portability (see table 6).
|Total amount||Amount by $2 payers||Rate of increase|
|Initial amount||$68 988 080||$21 517 854||3.21|
|Revised amount||$122 102 632||$4 449 312||27.44|
The rate of increase in the revised amount and the higher contribution above the level of $2 in the fixed-fee questions indicates that respondents were, on reflection, prepared to pay higher fees for portability. Also, the average payment is much higher when people were asked the revised value question. Four times as many people (57 per cent) were prepared to pay between $3 and $50 relative to those who were willing to pay nothing (16 per cent). In contrast, twice as many respondents (41 per cent) were prepared to pay up to $8 relative to those prepared to pay nothing (19 per cent) when the fixed questions were asked.
In the final calculation of the cost-benefit of portability policy, it is necessary to consider two amounts:
- The amount of initial $2 fee and the fact that just over three-quarters of respondents (76.5 per cent) were prepared to pay such a fee (plus another 2.8 per cent were also prepared to pay it subject to whether or not they would go overseas).
- The amount delivered on the basis of open fee question and the steep rate of increase from the minimal payment where the majority of respondents were prepared to increase payment to a value between $3 and $5.
Both amounts will be discussed in the next chapter.
- 3.1 Percentage of pensioners going overseas regardless of portability
- 3.2 Costs of long-term portability
- 3.3 Cost of overseas pensions
- 3.4 Costs saved because of portability policy
- 3.5 Net cost/saving of portability policy
To calculate the total social cost of long-term portability of Age Pensions, two types of costs or savings were taken into account:
- 'Opportunity costs'13: the total cost of portability policy calculated as possible savings if the policy were not implemented.
- 'Alternative costs': costs that would have to be incurred if the policy were not implemented. For the purpose of this research the opportunity costs are considered as the real cost of portability policy and alternative costs are calculated as savings associated with portability.
The following items were calculated as costs:
- The cost of pensions paid overseas to people who would have left, regardless of portability.
- Benefits forgone, such as taxes, earnings on capital taken overseas, child care and other services provided by pensioners.
The following items were calculated as savings:
- The difference between average pension entitlement in Australia and overseas.
- The costs of community infrastructure, such as health care and welfare services.
All costs were calculated as opportunity costs, meaning that only the costs that could be saved and possibly redeployed were considered. The cost that could be saved is the amount that was spent on pension payments to people who went overseas because portability policy allowed their pensions to be paid while they were overseas. To identify this group it was necessary to estimate the percentage of customers who would have gone overseas regardless of portability policy and to deduct this group from the total number of pensioners overseas.
There are some savings that offset the cost of portability. For example, customers who leave the country do not use services provided to age pensioners and therefore create savings. Also, some of them have the rate of their overseas pension reduced.14
In the first section of this chapter we estimate the percentage of people who would have gone overseas regardless of portability policy. Later, we calculate the costs and savings of the policy and derive its net cost-benefit.
The percentage of pensioners going overseas regardless of portability policy is estimated on the basis of:
- the survey data;
- data on departures before and after the introduction of portability; and
- the number of overseas customers with substantial income.
Survey respondents intending to go overseas for more than six months or permanently were asked whether they would still go if their pension were stopped. Of those intending to go overseas for more than six months (but not permanently) only 11 per cent firmly stated that they would still go even if their pension were not paid overseas.
However, the responses of those intending to go overseas permanently were different. More than 20 per cent said that it was very likely they would still go even if their pension were to stop.
A number of respondents did not provide a firm response to the question on whether they would or would not go overseas if their pensions stopped. They selected likely or unlikely options. It is reasonable to assume that 50 per cent of respondents who opted for the 'somewhat likely' response would select a positive answer if they were given the opportunity to select from just two positions. Thus, the relevant figures are increased to 19.4 per cent for those intending to go overseas for more than six months and to 29.6 per cent for permanent departures.
The charts that follow illustrate the proportion of those likely to go overseas regardless of portability.
Figure 10: Proportion of respondents who would go permanently without their pensions
Figure 11: Proportion of respondents who would go for more than 6 months without their pensions
The survey did not include a question asking whether respondents would be prepared to go overseas for a short period of time if their pensions were not paid. This question was omitted on the assumption that the majority of age pensioners on tight budgets would not consider going on holidays if their pension were stopped. In such circumstances their holiday expenditure would at least double. Not only would they have to pay for the trip, but they would also need to fund ongoing commitments in Australia. Indeed, if the survey data for departures of longer than 6 months and permanent departures are compared, it is estimated that fewer than 6 per cent would go overseas for a short period of time without their pensions.
Figure 12: Estimation of the percentage of people going for a short time without portability*
* The percentage of people going for a particular period includes 50 per cent of those who stated that they are 'somewhat likely to go'.
Departures of age pensioners before and after introduction of portability
Pensioners who travelled overseas before 1973 had their payment terminated. However, customers who stayed overseas for periods of less than one year had their pension reinstated
upon return and were paid retrospectively for up to 30 weeks. This situation radically changed after the introduction of general portability.
Since 1973, age pensioners can be paid their pensions overseas. By observing the change in departures data for age pensioners that occurred before and after the introduction of portability, we can estimate the percentage of customers who would have gone overseas regardless of portability.
We assumed that people who went overseas before the introduction of portability did so without relying on the payment of their age pensions. After 1973, age pensioners enjoyed continuous payment. Therefore, the difference between the number of age pensioners going before and after 1973 should indicate the overall behavioural effect of portability policy.
The available data on termination of pensions before 1973 do not distinguish between short- term, long-term and permanent absences. Nevertheless, on the basis of historical data it can be assumed that long-term and permanent departures together constitute around 10 per cent of departures. This figure is also consistent with the survey data where departures longer than 6 months amounted to 3.6 per cent (very likely) and 7.9 per cent (somewhat likely).
From a comparison of pre- and post-1973 departures it is clear that the introduction of portability policy increased the proportion of age pensioners among the group of people over 60 years of age leaving the country.15 In the five years before portability was introduced (from 1968 to 1972), the proportion of age pensioners in this group averaged around 2 per cent. After the introduction of portability (from 1975 to 1987) this proportion increased sevenfold, to an average of 15 per cent.16 The pre-portability proportion of 2 per cent was applied to post-1973 data on departures of aged persons (60 years and over) in order to arrive at an estimate of how many overseas pensioners leaving the country after the introduction of general portability would have gone regardless of portability. This calculation suggests that around 18 per cent of autonomous pensioners would have gone regardless of portability policy.
Figure 13: Proportion of those who would go for more than 6 months without their pensions
Overseas pensioners with substantial income
Another way of estimating the percentage of age pensioners going overseas regardless of portability of their pensions is by examining the number of pensioners overseas whose foreign income makes up a significant proportion (over 60 per cent) of their total income. It can be reasonably assumed that such pensioners would go regardless of portability since their high overseas pensions would partly compensate for the loss of their Australian pension.
The current data on overseas pensioners show that around 10 per cent of autonomous pensioners overseas have foreign income of more than 60 per cent of total income.17 Again, this figure is within the range of other figures discussed in this chapter.
Following the above analysis, it is estimated that between 10 per cent and 30 per cent of pensioners overseas would have gone overseas regardless of portability. To have a better understanding of these percentages, we distinguish between the data delivered from pensioners' intentions and actual figures.
|Comparison of data based on intentions and facts*|
|People intending to go for more than 6 months||19.40%|
|People intending to go permanently||29.60%|
|Departures pre- and post-1973||18.00%|
|Departures with substantial income—1999||10.00%|
|Average of people intending to go||24.50%|
|Average of people actually going||14.00%|
* These data include long-term and permanent departures
Our estimate of the percentage (5.53 per cent—see Figure 12) of the survey respondents who would go overseas for a short period of time without portability of their pensions is another important consideration.
From the estimated percentages of people going overseas regardless of portability, three scenarios can be explored:
- The 'best-case scenario', where only 14 per cent of pensioners overseas would go without portability of their pensions.
- The 'worst-case scenario', where 24.5 per cent of pensioners intending to go still go, even if their pensions were not paid during the absence.
- The 'most likely' scenario, where, on average, 19.25 per cent would go without portability of their pensions.
In the context of these scenarios, the cost of portability policy is calculated using 1998-99 data for 18 417 autonomous18 age pensioners overseas. For all calculations we use the following symbols:
PO — the total number of autonomous age pensioners overseas
PR — the percentage of pensioners who would have gone regardless of portability
The overseas pension rate is, on average, lower than the domestic rate because of the proportionalisation of the rate.19 In addition, ancillary benefits payable to pensioners in Australia such as rent assistance, pharmaceutical allowance and remote area allowance were not paid overseas until 20 September 2000. The average overseas entitlement is consistently lower than domestic entitlement and the difference has increased in recent years. In 1998-99, the average domestic rate for Age Pension was $8 157 a year compared to $7 500 for pensioners overseas.
Figure 14 compares rates over the last 5 years (in 1998-99 prices).
Figure 14: Comparison of overseas and domestic average rates
The cost was calculated according to the formula:
Cost = (average overseas rate x PO) - average domestic rate x (PO - (PO x PR)). Depending on
PR (which varies by scenario) the results are as in the table 8.
|14.00%||$ 8 929 983|
|24.50%||$24 703 829|
|19.25%||$16 816 906|
The administrative cost of portability policy comprises two elements:
- a possible saving on the administrative cost per pensioner in Australia for those who would have gone regardless of portability (i.e. 14 per cent or 19.25 per cent or 24.5 per cent of overseas pensioners multiplied by the average administrative cost per pensioner in Australia); and
- a possible saving on the difference between the administrative cost per pensioner overseas and the administrative cost per pensioner in Australia multiplied by the number of overseas pensioners who went overseas because of portability.
Due to the non-availability of data relating to the first element, the administrative costs were calculated on the basis of the second element only but the population was increased to include all autonomous pensioners overseas. This 'simplification' is, however, not considered critical for the final cost-benefit outcome.
|14.00%||$2 173 206|
|24.50%||$2 173 206|
|19.25%||$2 173 206|
Marginal social cost of raising public funds
Additional tax revenues are required to meet the nominal costs of most public projects or policies. However, raising taxes normally involves additional costs in the form of administrative costs and costs related to market distortions. These costs represent the so-called 'deadweight loss'. Therefore, the social opportunity cost of a public policy is measured by the sum of its nominal cost plus its deadweight loss. In the absence of capital rationing, the benefit/cost rule is that implementation of the policy should go ahead if its benefits exceed its social opportunity cost.
The estimate of the marginal social cost of raising public funds in Australia ia a coefficient within the range of 1-1.25 (Bond & Campbell 1997, p. 22). Therefore, in order to include the marginal social cost of raising public funds in the cost-benefit analysis, the benefits of portability policy should be higher than nominal costs increased by a factor in the range 1 to 1.25, thus:
B > NC x (1-1.25)
The nominal cost of portability policy includes only items for which the Government has to raise additional funds (for example, pensions that would otherwise stop and the administrative cost of paying pensions to those who would otherwise not be entitled). Of course, we would need to deduct those benefits of portability policy that directly reduce government outlays such as savings on community infrastructure costs. The final calculation is shown in Chapter 4,'Net cost-benefit of portability policy'. Since savings arising from the so-called 'alternative costs' override the direct nominal costs of portability policy (pension payments, running costs and forgone taxes), the marginal social cost of raising public funds was not included in the costs of portability.
According to taxation statistics for 1995-96, an average age pensioner paid around $281.40 a year in income tax.20 For the purposes of our calculations we have assumed that the average rate of taxation would remain the same in 1999 and in the years that follow. In addition, overseas pensioners would have also paid some indirect taxes and government fees. On average, in Australia, every adult paid around $1 000 in sales taxes a year.
Therefore, we have assumed that every autonomous pensioner overseas would have paid around $1 300 taxes a year in Australia.21
Cost = (average taxation rate) x (PO - PO x PR)
|14.00%||$ 20 590 206|
|24.50%||$ 18 076 286|
|19.25%||$ 19 333 246|
Forgone earnings on capital taken overseas
According to Centrelink Information Services data, the total value of assets of autonomous pensioners overseas as of August 1999 was $292.42 million. This is an average of $15 880 per pensioner. If the deeming rate of 4 per cent were applied, these assets would likely earn the following amounts in Australia.
Cost = (average value of assets per pensioner x 4 per cent) x (PO - PO x PR)
|14.00%||$ 10 057 524|
|24.50%||$ 8 829 570|
|19.25%||$ 9 443 547|
Community services provided by pensioners
According to ABS survey data, 15.25 per cent of Australians over 65 years of age spend an average of 164.4 hours per year in community work (ABS 1998). If a monetary value of $10 per hour of community work were applied, the total value of lost community benefits for autonomous pensioners who went overseas because of portability policy would be as follows.
Cost = 164.4 x $10 x [15.25 per cent x (PO - PO x PR)]
|14.00%||$ 3 970 900|
|24.50%||$ 3 486 081|
|19.25%||$ 3 728 491|
The total cost of portability policy is calculated by summing all the amounts from tables 8 to 12.
|14.00%||$ 45 721 820|
|24.50%||$ 57 268 972|
|19.25%||$ 51 495 396|
Some additional costs would have to be incurred if portability policy did not exist. These costs are generally described in cost-benefit analysis literature as 'alternative costs'. In the case of portability policy, these are generally related to spending on community infrastructure that supports age pensioners. In calculating those savings, the same approach is applied as in the previous section.
Health services costs
The most important item among so-called alternative costs is the saving on health expenditure. On average, health expenditure (including nursing homes) per person aged 65 years and over in Australia was $4 919 in 1993-94. Australian governments (Commonwealth and State) funded around 67 per cent of total costs, which means that government health expenditure per person aged 65 years and over was $3 296 in 1993-94 (in 1993-94 prices). Because the figure for 1998/99 was not available, this amount was translated into a 1998-99 figure using the CPI factor for each of the years22 and the growth rate per aged person of 2.8 per cent a year.23
The amount so derived was $4 217, which represents the government's average health cost per person aged 65 years and over in 1998-99. When this amount is multiplied by the number of autonomous pensioners who went overseas because of portability, we derive the following total health saving in 1998-99.
Cost = government's average health cost per aged person x (PO - PO x PR)
|14.00%||$ 66 791 461|
|24.50%||$ 58 636 689|
|19.25%||$ 62 714 075|
Pensioners with less than 25 years of Australian residence between the age of 16 years and Age Pension age are usually paid a reduced rate if they stay overseas for more than a year. Pensioners, who would have remained in Australia because of the non-portability of their pensions, would be paid the full pension rate. The difference between the full rate and the overseas rate would constitute an alternative cost. We did not calculate this alternative cost here because we considered the average overseas pension rate in calculating the portability costs.
The total government expenditure on welfare services (other than income support, nursing and residential care) per aged or disabled person was on average around $1 112 in 1995-96 (in 1995-96 prices) (AIHW 1997). To arrive at the figure for 1998-99 ($1 155) the appropriate CPIs were applied.
Cost = government's average welfare cost per aged person x (PO - PO x PR)
|14.00%||$ 18 293 606|
|24.50%||$ 16 060 084|
|19.25%||$ 17 176 845|
The total alternative costs (table 14 and table 15) are as follows.
|14.00%||$ 85 085 067|
|24.50%||$ 74 696 774|
|19.25%||$ 79 890 920|
The net saving produced by portability policy is calculated by subtracting the figures in table 13 from the figures in table 16, as shown in table 17.
|14.00%||$39 363 247|
|24.50%||$17 427 802|
|19.25%||$28 395 524|
Table 17 shows that portability policy produced significant net savings in Commonwealth outlays in 1998-99 under all scenarios. On the basis of historical data the following charts present the growth of net saving in constant (1998-99) prices with the GST effect taken into account from 2000-01. The values are in millions of dollars.
Figure 15: Costs and benefits with 14 per cent assumption ($ 000)
Figure 16: Costs and benefits with 19.25 per cent assumption ($ 000)
Figure 17: Costs and benefits with 24.5 per cent assumption ($ 000)
However, in the final calculation of the total net cost-benefit, the cost associated with a small percentage of people who would go overseas for a short time, for example on holidays, even if their pension stopped has to be taken into account. To calculate this cost the total number of autonomous pensioners overseas (18 417) is multiplied by 5.53 per cent of people who would go without their pensions and by an average domestic entitlement for Age Pension of $8 157 a year. Other costs associated with short-term portability are negligible.
The total cost of short-term portability is $8 307 579.04. This cost should be subtracted from the total net cost of portability policy. Therefore, the final savings are:
|14.00%||$31 055 668|
|24.50%||$9 120 223|
|19.25%||$20 087 945|
Future trend of the cost-benefit curve: determining (risk) factors
The above projection was based on the following assumptions:
- The number of overseas pensioners would continue to grow at the same rate as in the period 1994-98 (8.4 per cent).
- The real health cost per aged person would continue to grow at the same rate as in the period 1982-83 to 1994-95 (2.8 per cent a year).
- The gap between average domestic and average overseas rate would continue to increase at the same rate as in the period 1994-98.
- No other variable would change in real terms.
This picture, while simplified, is thought to be close to reality. Nevertheless, risk factors must still be taken into account.
Figure 18: Net benefit of the portability policy (likely scenario)
On the cost side, the overall taxation burden per age pensioner may either increase or decrease depending on the final outcome of the taxation reforms. However, it is more likely that it will decrease, at least in the short term, thus contributing to a reduction in the overall costs of portability policy. Centrelink administration costs for customers overseas may decrease over time due to system improvements (for example, direct credit rather than payment by cheque) or through other productivity increases. These changes would further reduce the costs of portability policy. It is unlikely that real interest rates will rise significantly given the level of competition in the financial market. Community services provided by aged persons may increase over time because the older population is becoming healthier and more active, but not enough to affect the forgone community benefits significantly, at least in the short term.
On the benefits side, savings on health costs may not continue to rise at the same rate as in the 1980s and 1990s. Productivity in the health sector may well increase and the Government's cost-reduction strategy in this sector may keep the real health cost per aged person steady in the future. However, it is highly unlikely that these costs will fall. It is also unlikely that the difference between the domestic rate and overseas rate will decrease. The difference may not grow at the same rate as it has in the past but it is likely to stay at current levels. Proportionalisation of the Australian overseas pension rate and globalisation trends (people less likely to spend most of their working life in one country) will continue to contribute to that difference. Real costs of welfare services per aged person are assumed to remain constant in the future. These costs are more likely to rise thereby contributing to an increase in the savings from portability policy.
On balance, the projected rise in net savings from portability policy may not be at the rate projected. However, it is thought more likely that net savings from portability policy will increase rather than decrease over time.
- 4.1 Calculation based on initial amount
- 4.2 Calculation based on revised amount
- 4.3 Calculation based on the rate of increase
- 4.4 Concluding remarks
In chapter 2, on the basis of the contingent value survey, the social benefit of portability policy was estimated. In chapter 3 the cost of portability policy, including some external costs, was calculated. In this chapter the social benefit and social costs will be compared. This comparison will allow us to make an assessment of the social and financial effectiveness of portability policy.
Before we compare the social benefit and costs of portability policy it is important to note that the contingent value survey data were collected on the basis of a hypothetical scenario. As previously discussed, the portability policy valuation of some respondents may well have been influenced by various hypothetical or structural biases. Our analysis has shown that these biases have only a marginal effect on the overall declared amount.
The major requirement of the cost-benefit methodology is that, in the final analysis, two figures have to be compared:
- the total amount of money declared by survey respondents, multiplied by the coefficient of Australian population; and,
- the total cost of portability policy, including some additional costs.
As discussed in chapter 3, the total cost of portability policy may be assessed under three scenarios. These scenarios depended on the estimated number of people who would go overseas regardless of whether their pensions would be portable. We will also consider three important social benefit amounts declared by the survey respondents:
- the amount declared on the basis of set fee questions;
- the amounts declared on the basis of revised value questions; and
- the rate of increase of payments from the minimum to the highest fee.
Costs are always expressed in monetary terms. Under the cost-benefit methodology the value of non-commodity goods can also be expressed in monetary terms as long as they can be privately appropriated and exclusively enjoyed goods (Anderson 1993, p. 193; Zerbe 1998, p. 424). Although the cost and social values are to be compared it has to be stressed that the expression of costs in monetary terms is always more real, being based on hard data, than the monetary value of the social benefit. However, it should be noted that, in both calculating the costs and evaluating the social benefit of portability policy, the same assumption was made. All analysis and valuations were made as if portability policy was not in existence. Both values therefore have the same 'abstract' characteristic. The assumption of the non-existence of portability policy proved valuable in evaluating a range of policy options. For example: what sort of portability would attract the strongest social support and which type of portability would be the most cost-efficient.
In the following section we will discuss the different scenarios and evaluate the total net cost-benefit of portability policy. In the final analysis we will apply a simple equation:
total social benefit - total costs* = net cost-benefit
* if costs are savings, as in the case of this report, they are expressed as (-) costs
The results will be interpreted by following the assumption:
- if net cost-benefit > 0 benefit
- if net cost-benefit = 0 neutral
- if net cost-benefit < 0 cost
The initial amount was calculated on the basis of the fixed fee questions (table 4). Two issues have to be considered in relation to this amount. More than 75 per cent of all respondents were prepared to pay the initial $2 fee (implying a total of $21 517 854). However, over 80 per cent of respondents were prepared to pay more. This means that the $2 fee does not reflect a real value of portability for a simple reason: most respondents were prepared to pay a $4 fee.
If the total initial amount is considered under all three scenarios the net cost-benefit value is positive.
|Scenario||Total social benefit||Total savings||Net benefit|
|1 - 14.00%||$68 988 080||$31 055 668||$100 043 748|
|2 - 24.50%||$68 988 080||$9 120 223||$78 108 303|
|3 - 19.25%||$68 988 080||$20 087 945||$89 076 025|
Social benefit is, on average, more than three times higher than savings.
Figure 19: Calculation based on initial amount
When considering the initial declared amount it is worth noting that respondents were asked to commit to set fees. Were the Government to introduce payment for portability, the gradation of fees could be perceived as a test of the respondents' level of acceptance. This may explain why most respondents accepted the $2 fee but were reluctant to pay fees above $4 mark.
To gain a better picture of the benefit of portability policy, we will assume an average fee of $3 and the number of respondents willing to increase the fee from $2 to $4. In this way, we can derive the total amount paid by those prepared to pay a basic fee and who were willing to increase it. This figure reflects the surplus value of portability policy.
|Scenario||Total social benefit||Total savings||Net benefit|
|1 - 14.00%||$26 039 115||$31 055 668||$57 094 783|
|2 - 24.50%||$26 039 115||$9 120 223||$35 159 338|
|3 - 19.25%||$26 039 115||$20 087 945||$46 127 060|
From the following chart it is clear that the minimum net social benefit based on the declaration of a $2 amount is around half of the total social benefit derived on the basis of the fixed fee questions.
Figure 20: Surplus value
The revised amount was calculated on the basis of the 'revised amount' question (table 3). Nearly three-quarters of respondents (72 per cent) were prepared to pay amounts between $2 and $50. The total amount delivered from the open value question is around twice as high as the amount declared on the basis of the 'set fee amount' questions.
|Scenario||Total social benefit||Total savings||Net benefit|
|1 - 14.00%||$122 102 632||$31 055 668||$153 158 300|
|2 - 24.50%||$122 102 632||$9 120 223||$131 222 855|
|3 - 19.25%||$122 102 632||$20 087 945||$142 190 577|
When the revised amount is considered, the social benefit is, on average, more than five times higher than savings.
Figure 21: Calculation based on initial amount
This revised amount, rather than the initial amount, indicates the true value that respondents attach to having their pension portable. Respondents were asked to indicate how much money they would be prepared to pay. This question followed the fixed fee questions and, because of that there could be some osmosis effect. In the fixed fee set of questions respondents were asked to declare a dollar valuation from a series of increasing amounts. At the free valuation stage declarations made by respondents may have been influenced by the previous set of values. For example, they may have remembered a figure from the previous set of questions and automatically repeated it without due consideration. Nevertheless, this amount reflects the social benefit of portability policy more accurately than that based on the fixed fee questions. The survey referred to a hypothetical market situation and the revised value resembled auction-style bidding—if I pay more then the payment of my pension is more likely. Therefore, the free valuation question more accurately reflects how much a person was prepared to pay for portability.
The highest percentage of respondents was prepared to pay a fee of between $3 and $5. Again, if we consider this as the surplus value of portability policy, it is clear that it is different to the pattern of payments observed in response to the set fee questions.
|Scenario||Total social benefit||Total savings||Net benefit|
|1 - 14.00%||$13 967 459||$31 055 668||$45 023 127|
|2 - 24.50%||$13 967 459||$9 120 223||$23 087 682|
|3 - 19.25%||$13 967 459||$20 087 945||$34 055 404|
From the following chart, we can see that minimal net benefit derived from the open value question is about one quarter the size of the total net benefit.
Figure 22: Surplus value
The rate of increase shows how much the total declared amount progressed from the basic $2 fee. If we assume that a $2 fee would secure pension portability, any increase in the amount that respondents were prepared to pay indicates the additional value of portability policy.
Figure 23: Additional value of portability policy
It is evident from the chart that the total amount delivered from the set fee questions is about half the amount delivered on the basis of the open question. This means that in an open market portability policy is valued more highly than when a fee is set. Also, the rate of increase is much higher when the respondents are given the opportunity to value the policy freely.
However, when the minimal net social benefit is derived from the open value questions and is compared with the amount derived from the set fee questions, the former is lower than the latter.
Figure 24: Comparison of minimal amounts
This outcome is mainly due to the wider spread of extremes in the amounts contributed on the basis of the open value questions compared with the set fee questions. For example, only one quarter (26 per cent) of respondents progressed to a $12 fee on the basis of the set fee questions, whereas more than 15 per cent were prepared to pay between $11 and $50 on the basis of the open value question.
From the ratio of increase shown on the chart, we can conclude that, if 24.5 per cent of people had gone overseas regardless of portability, the difference between minimal net benefits derived from the set fee questions and derived from the open value question is at a maximum.
Our analysis has shown that portability policy is of unquestionable benefit to society. The following chart graphically illustrates not only that savings are made, but that the average social benefit is close to five times higher than savings. In this context, even if we consider some shortcomings of the cost-benefit method, and the evaluation of social benefit and the total costs of the policy, it is indisputable that, overall, portability is strongly appreciated by the general public. This is manifest in the following chart.
Figure 25: Average benefits and savings
When we examine the average social benefit of portability policy more closely, it is clear that this benefit is not equally distributed. The strongest support is for short-term portability (up to 6 months of overseas absence), with only a small percentage of people not supporting short-term portability. In contrast, a significant number of respondents do not favour long-term (including permanent) portability of Australian pensions.
At the same time, more respondents intending to go overseas long-term or permanently were prepared to go if their pensions were stopped. If this decisiveness of long-term and permanent travellers is combined with the lack of support for long-term or permanent portability, then the social benefit of the policy sharply diminishes.
The following chart confirms that the highest support is for short-term portability. This support is reflected in the high declared amount for short-term portability and by opposition to long- term portability.
Figure 26: Contributions
On the other hand, short-term portability produces costs arising from payments to those who would have gone overseas short-term, regardless of portability. However, these costs are not significant, bearing in mind that a very small percentage of people are likely to travel short term, regardless of portability. Short-term portability does not produce savings because pensions are usually paid at the domestic rate and it does not have any effect on social infrastructure costs.
In contrast, long-term portability changes this situation dramatically in terms of savings. Savings arising from the lower average rate of overseas payment in comparison to the average domestic rate and savings from reduced costs on health and social welfare infrastructure significantly outweigh the cost of payments to the same number of pensioners, were they to decide to stay in the country.
The following chart shows how the social benefit and savings change in relation to short-term versus long-term portability. The total cost of short-term portability does not include savings to be had from reduced health or welfare costs because these services must still be maintained while pensioners are overseas short-term. It also does not include tax, earnings on capital taken overseas, or the loss of community services provided by pensioners.
Figure 27: Contributions, savings and total benefit
Given the substantial social benefit of short-term portability, we conclude that the only area where portability policy may be adjusted is in the area of long-term portability. Consideration could be given to technical legislative adjustments to increase the positive financial impact of long-term portability. Also, in the context of the criticism of portability policy sometimes seen in the mass media, raising the level of awareness in the community of the social and financial benefits of portability policy may be a worthwhile investment.
At the time the research was conducted the period of portability was payment specific. For example, Carer Payment was portable for up to 3 months and Wife Pension for up to 12 months. These discrepancies resulted from the piecemeal development of portability policy that had occurred since its introduction.
New portability rules were introduced on 20 September 2000. The new portability rules are consistent with the community attitude revealed in the research that for most people short- term portability means an overseas absence of up to 26 weeks. Under the new, standardised rules, all payments became portable for up to 26 weeks of temporary overseas absence. During the absence a customer must continue to qualify and there may therefore be some restrictions on portability. Many of the bureaucratic pre-departure procedures have been removed or simplified and the new portability rules mean that travelling is now far easier for many social security customers.
The questionnaire was placed on the Roy Morgan Research, Computer Assisted Telephone Interviewing (CATI) system. The CATI system is a quick, efficient and reliable way of conducting market research surveys. The questionnaire for a survey is programmed into the system. Interviewers are presented with the relevant questions and information as they proceed with an interview. All sequencing is automated through the software, so the possibility of sequencing mistakes and other forms of non-sampling errors are minimised. Where relevant, the CATI program automatically randomises or rotates pre-coded responses to questions. This eliminates a potential source of bias. As responses are entered directly into the system, the time required for data processing is considerably reduced.
Altogether 1 009 respondents were interviewed. This data was weighted to represent the Australian population, using one of Roy Morgan Research's standard weighting schemes that uses age, sex and area, and requires an 88 cell matrix (11x4x2) consisting of 11 areas, by 4 age groups, by sex, as shown below:
|NSW City||NSW x-City (incl.ACT)||18-24 yrs||Male|
|Victoria City||Victoria x-City||25-34 yrs||Female|
|Queensland City||Queensland x-City||35-49 yrs|
|South Australia City||South Aust.x-City (incl.NT)||50+ yrs|
|Western Australia City||Western Aust.x-city|
A randomly selected sample of households was drawn using the latest version of the Electronic White Pages. A maximum of four contact calls was made. Interviews were conducted among respondents who were aged 18 years or over. One interview was conducted per household. The following table shows the survey population and its relation to the Australian population.
|(popn. '000)||14 078||2 596||4 991||3 522||347||1 261||1 361|
|Male||6 918||1 284||2 447||1 720||170||620||677|
|Female uw||7 160||1 312||2 544||1 802||177||641||684|
|18-34 uw||4 723||894||1 671||1 200||74||410||474|
|35-49 uw||4 228||778||1 480||1 032||137||379||422|
|50-54 uw||1 279||210||452||358||29||86||146|
|ABOVE 65 uw||2 172||372||819||506||21||260||194|
* uw - unweighted
Contingent value questions
In order to estimate the value of pension portability, all respondents were presented with a hypothetical scenario, followed by a set of questions. Interpretation of the survey's results should be considered along with the precise wording of the scenario and these questions, which are:
I'd now like to read out an IMAGINARY SITUATION and ask your opinion. It is a made-up situation FOR RESEARCH PURPOSES ONLY.
Imagine that you are retired and receiving the Age Pension of $400 PER FORTNIGHT. This is your MAJOR SOURCE of income.
In order to have the right to receive your Age Pension when you are overseas, you would be required to pay a fee of $2 every fortnight \from when you retire \ now that you have retired. This would come to $52 for the year.
If you choose to pay the $2 fortnightly payment you continue to receive the Age Pension when you are overseas. If you choose NOT to pay the $2 fortnightly payment you would receive the Age Pension when in Australia but IT WOULD STOP when you are OVERSEAS.
Q13a. \When you retire \ now that you have retired \ would you choose to pay the $2 fortnightly fee for the right to CONTINUE TO RECEIVE THE AGE PENSION WHILE YOU WERE OVERSEAS, or would you choose NOT to make the payment?
IF SAID WOULD PAY $2 PER FORTNIGHT ON Q13a ASK,
Q 14a. Imagine the fee was $4 a fortnight, which would come to $104 for the year. Would you be prepared to pay $4 a fortnight?
IF SAID WOULD PAY $4 PER FORTNIGHT ON Q14a ASK:
Q14b. Imagine the fee was $8 a fortnight, which would come to $208 for the year. Would you be prepared to pay $8 per fortnight?
IF SAID WOULD PAY $8 PER FORTNIGHT ON Q14b ASK:
Q14c. Imagine the fee was $12 a fortnight, which would come to $312 for the year. Would you be prepared to pay $12 per fortnight?
Q15. What is the MAXIMUM amount you could AFFORD to pay and would be WILLING to pay from your $400 per fortnight to ensure your pension would continue to be paid to you if you were to go overseas?
Correlations between the maximal payment and demographic / travel variables
A comparison of the correlation coefficients for the different demographic variables shows that the correlation is highest for income (.081) i.e. income is the best predictor of the maximum amount willing to pay. Among the travel variables, the correlation is highest for the average number of weeks per year overseas (0.061). In both cases, the correlation coefficients are low, indicating that only a weak linear relationship exists.
Summary of regression analysis
The results from the regression analysis show that in spite of taking the variables with the highest correlation coefficients, the regressions have a very low value of R2 (percentage of the variance explained). The highest value of R2 is .011, (only 1.1 per cent of the variance explained), indicating that there is no clear linear relationship between the independent variables examined and the value of maximal payment.
Therefore, while the regression could be used to estimate the portability value for groups with various characteristics, the level of precision in such estimation would not be high.
5. As well as revising the long-standing Agreements with New Zealand (1 January 1995) and the United Kingdom (29 June 1992), new shared-responsibility Agreements have been implemented as follows: Italy from 1 September 1988, Canada from 1 September 1989, Spain from 3 June 1991, Malta from 1 July 1991, Ireland from 1 April 1992, The Netherlands from 1 April 1992, Portugal from 1 November 1992, Austria from 1 December 1992 and Cyprus from 1 January 1993.
6. 'Working life' occurs from the age of 16 years to Age Pension age. This gives a potential working life of 49 years for men and, with pension age for women currently set at 61 and a half years of age, 45 and a half years for women. The Age Pension age for women will increase over time to 65 years of age and will therefore bring working life residence in line with that of men.
8. A wife pensioner is an entitled person if she is a woman who is or was the partner of a man who was affected by the so-called 'Greek conspiracy' or a woman who was an Australian resident for at least ten years. A Widow B pensioner is an entitled person if she is a woman who is or was the partner of a man who was affected by the so-called 'Greek conspiracy', a woman who was an Australian resident for at least ten years, or a woman who is receiving a Widow B pension because her legal husband died.
12. If income between $35 000 and $39 999 is considered as an average income in Australia, the overseas-born group is represented in the same ratio as the general population. However, the overseas-born are over-represented at the extremes of low and high incomes.
13. Opportunity cost is defined as the value of the best alternatives or other opportunities forgone in obtaining an item or achieving an objective. Department of Finance 1997, Handbook of Cost-Benefit Analysis, p. 16.
15. To allow for the influence of emigration trends we weighted the permanent and long-term departure numbers for age pensioners by the total permanent and long-term departures of people aged over 60 yrs in the same periods.
19. Proportionalisation is used to calculate a rate of portable pensions for people with less than 25 years of residence in Australia between the age of 16 years and age pension age. The formula uses actual years of residence in Australia between 16 years and age pension age as the numerator and 25 years as the denominator to calculate the proportion of the Australian rate of pension to be paid overseas.
23. Average growth rate of Australian real health expenditure per person in the period 1998-99 was projected by using data from the period between 1982-83 and 1994-95. From 'Older Australia at a Glance'—a joint publication of the Australian Institute of Health and Welfare and the Office for Older Australians in the Commonwealth Department of Health and Aged Care.
Anderson, E. 1993, Value in Ethics and Economics, Harvard University Press, Cambridge, Mass.
Anderson, R. & Williams, A. 1975, Efficiency in the Social Services, Oxford University Press, London.
Australian Bureau of Statistics (ABS) 1998, Disability, Aging and Carers.
Australian Taxation Office, Taxation Statistics 1995-96, Individuals, Income from Government Pension by Age Group, Canberra.
Australia's Institute Health and Welfare 1997, Welfare Services Expenditure, Australia's Welfare Expenditure, Canberra.
Baldwin, R. 1995, Rules and Government, Clarendon Press, Oxford, p. 292.
Bond, K. A. & Campbell, H. F. 1997,'The cost of public funds in Australia', The Economic Record, vol. 73, no. 220, pp. 22-34.
Carson, R. T. & Mitchell, R. C. 1989,'Using surveys to value public goods: the contingent valuation method', Resources for the Future Publication, Washington DC.
Corne, R. & Sandler, T. 1986, The Theory of Externalities, Public Goods and Club Goods, Cambridge University Press, Cambridge.
Couanis, L. May 1993, Paying Australian pensions overseas. costs or savings?, unpublished research, thesis.
Cox, L. A., Cummings, R. G., & Freeman, A. M. 1986,'General methods for benefits assessment', Bentkover, J. D., Covello, V.T. & Mumpower, J. 1986, Benefits Assessments: The State of the Art, Kluwer, Dordrecht.
Department of Finance 1997, Handbook of Cost-Benefit Analysis, Canberra.
Dinwiddy, C. & Teal, F. 1996, Principles of Cost-Benefit Analysis for Developing Countries, Cambridge and New York.
Dmytrenko, A. 1997,'Cost benefit analysis', Records Management Quarterly, January, pp. 16-20.
Fink, A. (ed) 1995, The Survey Kit, vol. 1-9, SAGE Publications, London.
Galal, A., Jones, L. & Tandon, P. 1994, Welfare Consequences of Selling Public Enterprises: An Empirical Analysis, Washington.
Giardina, E. & Williams, A. (eds) 1993, Efficiency in the Public Sector: the Theory and Practice of Cost-Benefit Analysis, Edward Elgar, Aldershot.
Goodwin, N. R. Interdisciplinary Perspectives on Well-Being, www.tufts.edu/gdae
Human Resources Development Canada, Evaluation and Data Development 1998, Quasi- Experimental Evaluation.
Jones, L., Tandon, P. & Vogelsang, I. 1991, Selling Public Enterprises: A Cost Benefit Methodology, Cambridge (Mass.) and London.
Kalton, G & Moser, C. A. 1933, Survey Methods in Social Investigation, Dartmouth, Aldershot.
McNeill, P. 1990, Research Methods, Routledge, London.
Mansfield, C. 1998,'A consistent method for calibrating contingent value survey data', Southern Economic Journal, vol. 3, no. 64.
Nas, T. F. 1996, Cost-Benefit Analysis: Theory and Application, SAGE Publications, Thousand Oaks, London.
Ng,Y. K. 1993, The Necessity, Feasibility, and Adequacy of Cost-Benefit Analysis where Life or other 'Non-Economic' Factors are Concerned, Department of Economics, Monash University, seminar paper 11/1993.
Pearce, D. W. & Nash, C. A. 1981, The Social Appraisal of Projects, A Text in Cost-Benefit Analysis, McMillan Press, London.
Popay, J. & Williams, F. 1999,'Balancing polarities: framework for welfare research', Welfare Research. A Critical Review, Oakley, A., Popay, J. & Williams, F. (eds), University College London, London.
Rossi, P. H., Freeman, H. E. & Wright, S. R. 1999, Evaluation. A Systematic Approach. SAGE Publications, London.
Roy Morgan Research 1999, Survey on Portability Policy, unpublished research.
Singleton, R., Straits, B. C. & Straits, M. M. 1993, Approaches to Social Research, Oxford University Press, Oxford.
Wager, R. 1972, Care for the Elderly: An Exercise in Cost Benefit Analysis Commissioned by Essex (England) County Council, Institute of Municipal Treasures and Accountants, London.
Yew, K. N., 1993, The Necessity, Feasibility and Adequacy of Cost-Benefit Analysis where Life or other 'Non-Economic' Factors are Concerned, Melbourne.
Zerbe, R. O. 1998, 'Is Cost-benefit analysis legal? Three rules', Journal of Policy Analysis and Management, vol.17, no. 3.