Welfare fraud in Australia: Dimensions and issues

Foreword | Australia has a long tradition of providing welfare payments to vulnerable and disadvantaged citizens. The Australian Government introduced the aged pension in 1909 and the invalid pension in 1910. During World War II, and in the decades since, numerous additional benefits have been made available to a wide range of recipients. Almost one-third of Australians now receive some kind of direct welfare payment. For many, welfare provides a permanent, secure source of income. For others, government benefits are a vital stop-gap measure that contributes to equality of opportunity and longer term employment and career opportunities. However, the availability of welfare also creates temptations for fraud and allegations of fraud contribute to controversy about the appropriateness of welfare. This controversy can detract from public perceptions of the legitimacy of the welfare system and the dignity of welfare recipients. A major purpose of this paper is to help inform the debate about welfare by providing data on the size and dimensions of welfare fraud, including substantiated fraud, as evidenced by criminal convictions. Substantiated fraud represents a very small fraction of all welfare allocations, but losses can involve scams worth hundreds of thousands of dollars. Another Trends & Issues paper (Prenzler 2011) will examine current anti-fraud strategies and their impacts.

Adam Tomison

Director

This paper reports on key aspects of welfare fraud in Australia. It begins by outlining the basic aims of welfare systems that provide direct economic support, the vulnerability of these systems to fraud and issues around culpability and appropriate responses to suspected fraud. The paper also describes problems experienced when attempting to measure fraud and provides an analysis of available data about the size and dimensions of the problem, including case studies of major frauds. Overall, demonstrable fraud represents a very small fraction of all welfare transactions, but it also represents significant losses, demand for prosecution and loss recovery action. Overall, there is a need for greater consensus on the best ways to prevent fraud and deal with offenders.

The welfare state and the challenge of fraud

Welfare fraud—or ‘benefit’ or ‘social security’ fraud—is a controversial problem that has accompanied the growth of the welfare state. The modern welfare state developed in conjunction with the post World War II economic boom. It was designed, often in an ad hoc fashion, as a comprehensive system through which governments provide support for all citizens in need, with a view to eliminating poverty and enhancing health and wellbeing (McMahon 2005). Welfare systems frequently entail a wide range of living allowances paid to the elderly, unemployed, those with intellectual and physical disabilities, sole parents and students. Support also normally includes a range of partial, indirect or in-kind government funded benefits, such as child support payments and free or discounted medical services and childcare.

The welfare state has been the target of numerous criticisms. One standard critique is that it attracts fraud. There can certainly be little doubt that early benefit systems were highly vulnerable to abuse (Reeve 2006). It was not without justification that the terms ‘dole bludger’ or ‘welfare queen’ became part of the social and political discourse in many countries in the 1970s and 1980s.

Anecdotes about people feigning illness or disability, living on welfare while avoiding work, or collecting benefits while working, became a standard part of social gossip (Bradbury 1988). The right to apply for welfare and the availability of money created intrinsic temptations for people to attempt to obtain benefits fraudulently (Kuhlhorn 1997).

Welfare is usually organised around two main criteria—universal eligibility or means testing. Under universal eligibility, all persons fitting general criteria receive a benefit. For example, anyone over a specified age receives an old age pension. Conversely, means testing involves a second set of criteria related to income and assets. Recipients must meet a criterion, such as age, and also have income and assets below a specified threshold. Means testing is the primary form of welfare provision in Australia. It appears to be less costly, by reducing the number of recipients, and appears to be fairer in providing income only to those in genuine need. Alleged disadvantages of means testing include the requirement for a more complex bureaucracy and the creation of temptations for some applicants to understate or hide income and assets (Kuhlhorn 1997).

Something of the scope for fraud can be seen in statistics for Australia’s federal welfare agency Centrelink (located in the Department of Human Services portfolio). In 2008–09, Centrelink distributed approximately $86.6b to 6.8 million customers, including $10.4m in individual entitlements, across 140 benefit types on behalf of 27 government departments and agencies. It approved 2.7 million new claims, operated over 1,000 service delivery centres, employed just under 28,000 staff and made over six billion transactions on customer records (Centrelink 2009a: 28).

Ensuring payment integrity

A number of factors have led to reassessments of liberal access to welfare and a concern for ‘payment integrity’ (Centrelink 2009b; Reeve 2006). One factor was the strain on existing allocations caused by the continuing enlargement of entitlements. A second was the contraction of state resources under global recessions from oil crises and other shocks from the 1970s. Another was the rise of ‘economic rationalism’ and ‘user pays’ philosophies, associated with voter revolts against high spending, high tax and high debt governments. Media exposés of fraud also fuelled popular opinion against welfare cheats. An Australian public opinion survey in the mid 1980s found that social security fraud worth $1,000 was considered worse than tax evasion or medical fraud worth $5,000 (Wilson, Walker & Mukherjee 1986).

One effect of these developments was to focus attention on improving mechanisms for ensuring benefits went to genuine cases. ‘Modernisation’ of systems entailed better screening processes at the point of application for support, as well as closer scrutiny of existing welfare recipients to ensure they remained eligible (Green 2008). In Australia, the Fraser Government (1975–1983) and the Hawke–Keating Government (1983–1996) tightened compliance measures in a number of areas, including stricter work search tests for the unemployed and data-matching. The Howard Government (1996–2007) made combating fraud a major plank in its first victorious election campaign and expanded existing prevention and detection measures while introducing new initiatives (Dunlevy & Hannon 1997; Kingston 1996).

Innovations in combating welfare fraud

Centrelink was created in 1997 by the Howard Government as a ‘one-stop-shop’ for government social security services. This was primarily an efficiency measure in terms of service delivery. At the same time, the establishment of Centrelink allowed for the centralisation and standardisation of anti-fraud methods. The Centrelink system represented the introduction of a ‘purchaser/provider separation’, with arrangements between Centrelink and other departments based on a ‘business partnership’ agreement (Mulgan 2002).

Internationally, the last 30 years have seen considerable innovation in welfare systems, focused on identifying both ‘error’ and ‘fraud’. In the United Kingdom, for example, three categories are used to describe inaccuracies in benefits claimed or provided:

The Department (for Work and Pensions) defines fraud as those cases where customers deliberately claim money to which they are not entitled. Customer error occurs when customers provide information to the Department which is inaccurate, incomplete or untimely, but without dishonest intent, and as a result the benefit paid is inaccurate. Official error occurs when officials fail to apply specific rules or do not take into account all the notified circumstances (NAO 2008: 6).

Many of the innovations aimed at improving compliance have been driven by legislative requirements. Centrelink is subject to the Commonwealth Financial Management and Accountability Act 1997 and must comply with the Commonwealth Fraud Control Guidelines (AGD 2002). Fraud management reviews by the Australian National Audit Office (eg 2008a, 2007, 2006, 2001) have also driven change. Anti-fraud measures adopted overseas and in Australia include the following:

• data-matching between government agencies;
• stepped up identity verification checks;
• covert surveillance and video recording;
• stepped up investigations, with greater use of forensic accounting and site visits;
• increased prosecutions;
• increased recoveries through debt collection strategies and asset forfeiture;
• advertising rules and compliance requirements;
• deterrence through advertising prosecutions and convictions; and
• public tip-off lines (Prenzler 2011).

Measuring welfare fraud and the welfare debate

Many of these anti-fraud measures have been criticised as stigmatising welfare recipients and generating ‘a punitive approach to income support’ that is overly-reliant on criminal prosecutions (Bradbury 1988: 26). Administrative assessments and adjustments, it is claimed, are more efficient and provide for a potentially better resolution of disputed assessments (Freiberg 1989). A ‘get tough’ approach easily criminalises recipients who have made genuine errors in reporting their circumstances and this potential is exacerbated by the casualisation of labour and the shifting circumstances of the working poor and unemployed (Marston & Walsh 2008).

Welfare fraud is difficult to measure because it falls outside the two main crime measures of incidents reported to police and victim experience surveys. One recent attempt to measure welfare fraud, using international econometric benchmarks, was unspecific but nonetheless assumed high levels of fraud based on the relative generosity of the Australian system. Bajada (2005: 184) concluded that, ‘there appears to be a significant number of people in Australia fraudulently accepting welfare benefit payments while in receipt of subterranean income’. From a different perspective, Peter Saunders, from the Centre for Independent Studies, has argued that, although ‘the welfare lobby insists fraud is not a serious problem’, surveys of the unemployed show that up to 75 percent are not genuinely willing to search for or accept available jobs and are therefore, in a sense, ‘fraudulent’ (Saunders 2003: 11–12; see Colmar Brunton Social Research 2002: 20). In contrast, Michael Raper, President of the National Welfare Rights Network, using conviction rates, has argued there is very little fraud in social security—‘It’s pretty tight and hard already. Less than half of one percent of social security debt is fraud’ (in Karvelis 2008: 4).

Overseas findings

The UK Department for Work and Pensions estimated that in 2008–09, approximately 2.2 percent of all benefit expenditures, or £3b, was overpaid as a result of fraud and error (DWP 2009). Half of this, about £1.1b, was attributed to fraud, although this was based on a sampling procedure rather than convictions. The figure represented an increase, from a low of £0.6b in 2005–06, despite concerted efforts by the department to stop fraud (NAO 2008).

In the United States in 2008–09, the Social Security Administration Office of the Inspector General (2009) received 129,495 allegations of fraud and closed 8,065 cases, with 1,486 criminal prosecutions. These activities involved over US$2.9b in ‘questioned costs’; with US$23.3m in recoveries, US$2.8m in fines and a further US$25.5m in settlements, judgement and restitution orders.

Source: Centrelink 2010

Australian data

The following section presents data supplied by Centrelink on its compliance and fraud-related activities and outcomes. Unlike the UK Department for Work and Pensions, Centrelink does not provide estimates of fraud but reports on detected errors and fraud prosecution actions and outcomes. Formal fraud investigations are usually initiated through compliance and eligibility reviews. Reviews occur in large numbers each year. There is a crossover of triggers and methods, including routine data-matching, random sampling, identity checks and public tip offs.

Table 1 reports on the outcomes of reviews for the three year period 2006–07 to 2008–09. Of note is the fact that typically, only 15.7 percent of reviews led to cancellations or reductions in payments. Of these, as few as 0.8 percent were referred to the Commonwealth Director of Public Prosecutions (CDPP); with 0.5 percent being prosecuted. Prosecutions resulted in a 98.8 percent conviction rate. Overall, in the three years, 0.04 percent of customers were convicted of fraud. For the same period, fraud investigations were estimated to have produced $380.6m in gross savings and amounts targeted for recovery. This compares with $1.4b in overpayments identified and debts generated from the review process. Fraud therefore accounted for approximately 26.2 percent of invalid payments. Furthermore, on average, only 15.1 percent of investigations resulted in a prosecution referral. In 2008–09, Centrelink referrals accounted for 69 percent of defendants prosecuted by the CDPP (2009: 115–116).

Figure 1: Trends in compliance reviews, 1997–08 to 2008–09 (n) [see attached PDF for graph]

Table 2 provides a snapshot of fraud across the top 15 benefit types. Within this group, the Single Parenting Payment and Newstart Allowance (unemployment benefit) together accounted for 72 percent of convictions and $33.5m of debt. The Disability Support Pension and Partnered Parenting Payment together accounted for a further 14.7 percent and $7.6m of debt.

Figure 1 shows longer term trends for compliance reviews and adjustments for the 12 year period from 1997–2008 (when Centrelink was established) to 2008–09. They show that, in terms of the number of Centrelink customers, compliance reviews increased by 54.5 percent from an average of 41.1 percent of customers up to 2001–02, to an average 63.4 percent subsequently, while cancellations or adjustments more than doubled from 4.3 percent to 10.1 percent.

Figure 2 shows that referrals to the CDPP have increased less dramatically, with prosecutions and convictions at a fairly stable rate.

Source: Centrelink unpublished data 2010

Note: Cases can be recorded against more than 1 benefit type

Case studies

Case studies provide another source of information concerning welfare fraud. Media releases about major cases alert the public to the anti-fraud work of welfare departments and the costs of fraud. For example, a conviction in early 2008 reported in a media release involved benefit fraud of more than $195,000. The fraud was perpetrated through the creation of a false identity and continued for 22 years (Ludwig 2008b). Case studies are also reported in order to deter potential offenders. The recent case studies below, provided by Centrelink, show how scams can operate for many years and involve multiple identities, adding up to hundreds of thousands of dollars of fraudulent payments. The examples also show how Centrelink uses case study analysis as a learning tool to improve prevention.

Case study 1

A 52 year old male received a disability support pension under a fraudulent identity, created with fake documents, while also receiving disability support pension under his legitimate identity. The offences extended over a 14 year period. They were discovered when a group of customers was examined for not using their Medicare cards during a five year period. After examination of customer files, it was established that there were two customers who shared many similarities including similar handwriting, similar medical histories and similar past addresses, and the same address was used for correspondence. As a result of this fraud, the offender incurred a debt of approximately $240,000. In response to this case, Centrelink introduced regular data-matching with Medicare to detect customers who have not used their Medicare cards for five years.

Case study 2

A 35 year old government employee fraudulently created 23 Baby Bonus claims involving 58 fictitious children. The offences occurred over a four month period. As a result of the fraud, the offender incurred a debt of $318,286. The offender targeted age pension recipients and accessed customer records to obtain the tax file numbers of recently deceased customers. He then created new customers and granted claims for Baby Bonus and Maternity Immunisation Allowance payments. In many of the claims, the children were registered as stillborn, so additional payments of Bereavement Allowance were paid. The fraud was discovered by internal controls that detect suspicious access to customer records, as well as by ‘identity scoring controls’ that detect fabricated identities. As a result of this case, system enhancements were implemented to prevent the use of encrypted tax file numbers in this unauthorised manner.

Case study 3

A 35 year old female fraudulently claimed the Australian Government Disaster Relief Payment and the Recovery Subsidy Assistance. A total of 21 claims were lodged. As a result of the fraud, the customer incurred a debt of $47,925. The offences occurred over a three week period. The customer took advantage of the proof of identity protocols that were relaxed to assist disaster victims. She fabricated identities and manufactured circumstances to meet disaster relief eligibility rules. The fraud was discovered through claim analysis, where similarities in names were identified and in some cases, common destination bank accounts were used. As a result of this case, standard analysis rules were developed and implemented, and these are now run against all relief payments after disasters.

Figure 2: Trends in prosecution referrals and outcomes, 1997–08 to 2008–09 (%) [see attached PDF for graph]

Discussion

The first sections of this paper introduced some themes in the debate about welfare fraud. There is a lobby that sees anti-fraud measures as overly punitive, with recipients who make mistakes being criminalised and driven further into debt by recovery orders. By contrast, there is also a lobby that argues that welfare is too easy to obtain, that it attracts fraud and reduces government spending in areas of general welfare such as health and education.

Both sides in the debate tend to agree that public opinion on the topic is important and that the fair delivery of services to disadvantaged persons should be a consensus public policy position (Bajada 2005; Green 2008). Finding common ground on fair and effective strategies for reducing fraud and dealing with non-compliance is therefore a potentially important means of enhancing trust in the system.

The data outlined above provide some openings into this debate. It can be seen that in Australia, prosecution referrals for welfare fraud account for a small fraction of all assessments and that on average, only 0.04 percent of the 6.5 million plus welfare recipients are convicted of fraud each year. The system of referrals to the CDPP also ensures cases go through an independent filtering process. Only the strongest cases are pursued, as indicated by the very high conviction rates achieved in the tough arena of the criminal courts—with a standard of evidence ‘beyond reasonable doubt’. Analysis of the types of serious fraud cases outlined above shows that these are usually carefully planned with clear criminal intent (Webb 2001). The Commonwealth Fraud Control Guidelines include an obligation to prosecute offenders (AGD 2002), subject to consideration of mitigating circumstances and a number of other factors set out in the Prosecution Policy of the Commonwealth (AGD 2008). However, the apparent high threshold for welfare fraud prosecutions means that cases potentially involving fraud might be categorised as ‘error’. It should also be kept in mind that most welfare agencies consistently ‘write off’ large amounts of debt. For example, Centrelink waived $67m in debt in 2008–09—in some cases because the pursuit of debtors was considered unwarranted (ANAO 2008b; Centrelink 2009b).

The official data reported in this study only take us so far. More detailed studies might assist in developing greater consensus about how to respond to fraud. For example, some preliminary research has been done on the range of losses involved in convicted fraud cases and the sentencing outcomes. Marston and Walsh (2008) studied 80 social security fraud cases in two Magistrates’ courts. They found that the average amount involved was just over $10,000. The largest amount was $30,105 and the lowest was $162. There were no cases of identity fraud or elaborate scams. In their view, the findings ‘challenge the stereotype of the organised criminal willingly defrauding the Commonwealth Government for large sums of money’ (Marston & Walsh 2008: 297). The researchers concluded that in many cases, it was plausible that circumstances pointed to error rather than criminal intent (see also Hughes 2008). In questioning the value of prosecuting many of these cases they also pointed to the fact that 85 percent of persons had already repaid all or some of the debt, were further burdened with court costs and that very low tariff penalties were imposed in almost all cases. Of 96 penalties, there were only two prison terms. The remainder involved good behaviour bonds (58%), community service orders (16%), suspended sentences (14%), fines (6%), or probation (3%). This study did not include higher courts where more serious cases are prosecuted. Nonetheless, the findings suggest there may be little value in pursuing minor matters in the criminal courts when administrative remedies are available.

A final issue concerns the preventive effects of anti-fraud measures. At present, ‘success’ against welfare fraud appears to lie primarily in the area of ‘secondary prevention’; that is, in detecting and stopping ongoing fraud after it has begun. With secondary prevention, the benefits obtained from halting future losses are enlarged by the recovery of past losses through repayment orders against convicted offenders. However, in terms of the overall picture, something of a paradoxical situation can be seen. As fraud prevention efforts increase, more suspected fraud is uncovered. The result is that there are few signs of substantive reductions in fraud and there is an ongoing ‘roll call’ of offenders convicted in the courts—approximately 3,000 each year (see Table 1). Consequently, the most significant challenge for welfare fraud policy is to make a more decisive shift from secondary prevention to primary prevention; that is, to prevent fraud occurring in the first place and reduce the need for expensive and difficult secondary level processes of detection, prosecution, punishment and restitution. This challenge is recognised by the Australian Government (Ludwig 2008a) and addressed in more detail in another Trends & Issues paper (Prenzler 2011).

Conclusion

The last two decades in Australia have witnessed a growing commitment by the Australian Government to combat welfare fraud. This is a crime problem that is difficult to measure, but available indicators suggest that fraud represents an ongoing threat to the integrity of welfare payments. The issue of how to respond to welfare fraud is also difficult and attracts ongoing controversy. It is possible, however, that greater consensus could be found through finding more effective primary prevention measures and making more use of administrative responses to lower level suspected fraud.

References

All URLs correct at May 2011

• Attorney-General’s Department (AGD) 2008. Prosecution policy of the Commonwealth. Canberra: AGD
• Attorney-General’s Department (AGD) 2002. Commonwealth fraud control guidelines. Canberra: AGD
• Australian National Audit Office (ANAO) 2008a. Centrelink’s tip-off system. Canberra: ANAO
• Australian National Audit Office (ANAO) 2008b. Management of customer debt—Follow-up audit (Centrelink). Canberra: ANAO
• Australian National Audit Office (ANAO) 2007. Proof of identity for accessing Centrelink payments. Canberra: ANAO
• Australian National Audit Office (ANAO) 2006. Assuring Centrelink payments—The role of the random sample survey programme. Canberra: ANAO
• Australian National Audit Office (ANAO) 2001. Management of fraud and incorrect payment in Centrelink. Canberra: ANAO
• Bajada C 2005. Unemployment and the underground economy in Australia. Applied Economics 37(2): 177–189
• Bradbury B 1988. Welfare fraud, work incentives and income support for the unemployed. Sydney: Social Welfare Research Centre, University of New South Wales
• Colmar Brunton Social Research 2002. Job seeker attitudinal segmentation. Canberra: Department of Employment & Workplace Relations
• Centrelink 2009a. Annual report, 2008–2009. Canberra: Centrelink
• Centrelink 2009b. Payment integrity information package, 2008–9. Canberra: Business Integrity and Debt Reporting Section, Centrelink
• Commonwealth Director of Public Prosecutions (CDPP) 2009. Annual report 2008–09. Canberra: CDPP
• Department for Work and Pensions (DWP) 2009. Fraud and error in the benefit system: April 2008 to March 2009. London: DWP
• Dunlevy S & Hannon K 1997. Red handed—330,000 welfare payments stopped. Daily Telegraph 30 October: 1
• Freiberg A 1989. Social security prosecutions and overpayment recovery. Australian and New Zealand Journal of Criminology 22(4): 213–236
• Green K 2008. Child poverty and welfare reform. Presentation to Brotherhood of St. Laurence. Melbourne: April 7
• Hughes C 2008. Caught in the safety net: The Costs of Centrelink debt recovery and prosecution. Hobart: Anglicare Tasmania
• Karvelis P 2008. Welfare fraud on Rudd’s hit list. The Australian January 25: 4
• Kingston M 1996. $19m in savings show little room for welfare crackdown. Sydney Morning Herald 21 August: 6
• Kuhlhorn E 1997. Housing allowances in a welfare society: Reducing the temptation to cheat, in Clarke R (ed), Situational crime prevention: Successful case studies. Monsey, NY: Criminal Justice Press
• Ludwig J 2008a. Address to smart cards summit. Ministerial speech 11 June. http://www.dhs.gov.au/media/archives/ludwig/speeches/080611-Ministerial-address-smart-cards-summit.php  
• Ludwig J 2008b. Three year prison for $195,000 welfare fraudster. Media release 9 February.  
• Marston G & Walsh T 2008. A case of misrepresentation: Social security fraud and the criminal justice system in Australia. Griffith Law Review 17(1): 285–300
• McMahon A 2005. Understanding the welfare state, in McMahon A, Thomson J & Williams C (eds), Understanding the Australian welfare state: Key documents and themes. Melbourne: Tertiary Press
• Mulgan R 2002. Public accountability of provider agencies: The case of the Australian ‘Centrelink’. International Review of Administrative Services 68(1): 45–59
• National Audit Office (NAO) 2008. Department for Work and Pensions: Progress in tackling benefit fraud. London: NAO
• Prenzler T 2011. Detecting and preventing welfare fraud. Trends & Issues in Crime and Criminal Justice no. 418. Canberra: Australian Institute of Criminology. https://www.aic.gov.au/publications/tandi/tandi418
• Reeve G 2006. Human services administration: A More intelligent approach to reducing benefit fraud. Public Administration Today January/March: 37–45
• Saunders P 2003. Help and hassle: Do people on welfare really want to work? Policy 19(2): 8–14
• Social Security Administration Office of the Inspector General (SSAOIG) 2009. Semi-annual report to Congress, April 1, 2009–September 30, 2009. Baltimore: SSAOIG
• Webb J 2001. Prosecuting social security fraud: Protecting the trust fund for present and future generations. High Tech and Investment Fraud 49(6): 1–5
• Wilson P, Walker J & Mukherjee S 1986. How the public sees crime. Trends & Issues in Crime and Criminal Justice no. 2. Canberra: Australian Institute of Criminology. https://www.aic.gov.au/publications/tandi/tandi2

About the author

Tim Prenzler is a Chief Investigator at the Griffith University Brisbane node of the Australian Research Council Centre of Excellence in Policing and Security.