Mandating the Measurement of Fraud

Legislating against Loss

M. Tunley

• Publisher: Palgrave Pivot
• Year: 2014
• ISBN:
  9781137406286
• Topic: Social Sciences
• Subtopic: Criminology
• Index: Social Sciences

This project examines the concept of fraud loss measurement by critiquing existing measurement methodologies, and argues for the mandating of fraud loss measurement by enforced self regulation, the creation of a British Standard of fraud loss measurement, and the establishment of an information exchange matrix to develop best practice.

1

The Issue

Abstract: Fraud is the most costly crime to society, with estimated losses exceeding those from other acquisitive crimes such as burglary and robbery. However, there have been limited attempts to accurately measure the extent and nature of these losses, which suggests that this figure is only the tip of the iceberg. This chapter introduces the issues addressed within this book by first offering examples of attempts to calculate losses before addressing the question ‘what is fraud?’ and evidencing that the accurate measurement of fraud is something that is occasionally aspired to but rarely achieved. Examples are then provided of how fraud is measured, and to offer a broader context, a brief review of some measures of corruption is also offered. The focus then moves on to the research argument, followed by a discussion on the value of this research. This chapter then outlines the research methodology, while also discussing the generalizabilty in terms of limitations and scope.

Tunley, Martin. Mandating the Measurement of Fraud: Legislating against Loss. Basingstoke, Palgrave Macmillan, 2014. DOI: 10.1057/9781137406286.0005.

How much does fraud cost?

There are varying estimates of the cost of fraud to the UK, ranging from £7 billion to £73 billion. These annual loss figures include the following: £6.8 billion to £13.8 billion (National Economic Research Associates [NERA], 2000), £16 billion (Norwich Union, 2005), £40 billion (RSM Robson Rodes, 2004) and £72 billion (Mishcon de Reya, 2005). Reviews of these data suggest that losses may range between £14 and £72 billion (Fraud Review Team, 2006; Levi, Burrows, Fleming and Hopkins, 2007). During 2010 annual fraud losses by the public sector, private sector and charities were estimated by the National Fraud Authority (NFA) to be £30.5 billion (NFA, 2010a, p. 8, 2010b, p. 1), increasing in 2012 to an estimated loss of £73 billion (NFA, 2012). Interestingly, the most recent estimation calculates total fraud losses to be £52 billion (NFA, 2013).

Clearly, not all these can be correct; consequently, with such a wide disparity of estimates of the cost of fraud, there is a need for an evaluation of existing measurement methodology to develop a more accurate mechanism which produces meaningful data.

What is fraud?

We have seen how there have been differing attempts to measure fraud and the extreme variations in the loss figures presented. It is therefore relevant to explore what actually constitutes fraud to help to illustrate why there are such wide discrepancies in loss figures. Historically, that is to say prior to the Fraud Act 2006, one of the most frequently asked questions was ‘what is fraud?’ The collective academic and practitioner response being that there was no ‘definitive’ or ‘universal’ definition (Doig, Johnson and Levi, 2001; Fraud Advisory Panel, 1999). This lack of a universal definition has significantly limited any meaningful measurement of fraud. The consequence being that organizations apply their own bespoke definition which restricts any meaningful analysis and comparability. The Fraud Advisory Panel’s (1999) study of published literature on fraud identifies only one report offering a definition of fraud (p. 6), this being ‘the use of deception with the intention of obtaining advantage, avoiding an obligation or causing loss to a third party’ (Her Majesty’s Treasury, 1995, p. 6). This definition, while being rather dated, does summarize the key elements of the Fraud Act 2006 and warrants consideration when developing a standard definition for measurement purposes, being both succinct and unambiguous.

The aim of the Fraud Act 2006 is to simplify matters and improve general understanding of this crime. According to the legislation fraud can be perpetrated in three clearly defined ways:

By false representation.

By abuse of position.

By failure to disclose.

The statute does offer a description of how fraud is committed; however, it fails to answer the definitional question of what actually constitutes fraud.

Optimistically, Hoare (2007) argues that this statute facilitates ‘effective measurement of fraud’ by making ‘recording and reporting fraud easier’ (p. 277). While being of relevance to this research, this view is based on the presumption that fraud losses may only be measured using reported or detected data, which is no longer the case.

Unfortunately, this legislation has failed to address the problem of multiple definitions of fraud, because ‘drafting for legal purposes seldom provides ... behavioural categorisation that corresponds to the way individuals and businesses categorise frauds’ (Levi et al., 2007, p. 9). This observation being evidenced by the continuing range of fraud definitions used for measurement purposes following the introduction of this statute (Audit Commission, 2013; Levi and Burrows, 2008). The latter defining fraud as ‘any intentional false representation, including failure to declare information or abuse of position that is carried out to make gain, cause loss or expose another to the risk of loss’ (Audit Commission, 2013, p. 8). This exemplifies two important common themes required within a universal definition, these being financial gain and causing loss.

The civil definition of fraud, Derry v Peek (1889), considers fraud to have been proved ‘when it is shown that a false representation has been made (a) knowingly, or (b) without belief in its truth, or (c) recklessly, careless whether it be true or false’. The burden of proof is based on the balance of probabilities rather than ‘beyond reasonable doubt’, thus if applied for measurement purposes would include cases where fraud is identified but with insufficient evidence for a criminal prosecution. Accordingly, it offers a potential option as a standard fraud definition for the purpose of more accurate loss measurement.

Why do we measure crime and fraud?

It is also worth considering why crime and fraud are actually measured. Foucault (1977, 1979, 2000) argues that collecting information about individuals forms part of a government strategy to extend control over the population. Levi and Burrows (2008, p. 293) make a similar observation, suggesting that the collection of crime data ‘to serve the panoptican poses a question, namely answers are required concerning what is required and what is not collected by those managing the state’. This suggests there may be a political agenda in terms of data collection, and explanations are required as to why on occasions the state fails to look too closely at certain crime types. Brand and Price (2000, p. 3) offer a simple explanation for the collection of crime data, suggesting that it provides a way of measuring crime reduction policies. Accordingly, a precise representation of fraud losses enables focused investment in, and deployment of, any tactical counter-fraud resource.

Do we really look for fraud?

It has been suggested that policies encouraging individuals to report fraud may result in an unachievable public expectation on law enforcement agencies to address this issue (Levi and Burrows, 2008, p. 315). To alleviate such a risk, it may be in the interest of law enforcement agencies and the government to undercount fraud, which may explain limited interest in fraud loss measurement.

It is therefore worthwhile considering why these data are collected, and possible motives for not looking too hard. The Home Office has been criticized for targeting research to suit the government’s political agenda (Walters, 2005, p. 6). This charge may also be levelled at the collection, or in some cases lack of collection, by central government departments of fraud loss data. In support of this contention, this book offers the findings of a public sector fraud survey which reveals that in the preceding 12 months, only 52 per cent of government-owned enterprises reported economic crime (PriceWaterhouseCoopers, 2010).

Insurance companies pose another dilemma, frequently absorbing losses because fraud is seen as a consequence of a high volume of transactions. This suggests that fraud is considered a business cost by these organizations, and therefore should be measured accurately. Reluctance to confront this issue by the private sector due to fear of organizational embarrassment, or in the case of the charitable sector, concerns that exposure may impact on donations and may explain the limited engagement with fraud loss measurement by these sectors. Limited appreciation of the amount of potential losses to fraud may also result in these organizations believing it may be more cost effective to ignore rather than address the issue.

How do we measure fraud?

There are different mechanisms for measuring fraud used by both the public and private sectors offering varying levels of accuracy and statistical confidence. These include:

(Censuses drawing on) administrative records of fraud reports.

Probability and non-probability sample surveys of individuals and firms as fraud victims.

Audits of probability samples of customers/accounts/transactions/expenditures to uncover fraud losses.

Analyses of samples of Suspicious Activity Reports filed on suspicion of money laundering.

Analyses of samples of offenders convicted of certain frauds or of law enforcement case information.

(Fleming, 2009, p. 11)

This array of methods explains why there is a range of estimates of the true cost of fraud, and evidences an urgent need to apply a more consistent approach to loss measurement.

An additional factor impacting upon the calculation of fraud losses and offering an explanation for the variety of measures is cost. The amount of resources devoted to measurement exercises influences the reliability and statistical confidence of resultant data. Limited resources may result in sporadic measurement exercises, with insufficient samples which subsequently generate unreliable data that offers limited scope for developing cost-effective control strategies. To address this problem there is a pressing requirement to change the way in which fraud losses are viewed. They must be treated as a business cost and measured accordingly.

Measuring corruption

There is sometimes a blurring of the edges between fraud and corruption. For example, fraud by those holding a position of trust may also constitute corruption. The measurement of fraud and corruption do pose different challenges (Samford, Shacklock, Connors and Galtung, 2006); however there is some overlap. Accordingly, it is worth reviewing some measures of corruption to offer a broader contextualization of the measurement of crimes that are covert by nature. There are a wide range of corruption measures worldwide, and this chapter confines itself to a small sample to illustrate that, similar to fraud, differing measurement methodologies exist. Furthermore, similar to fraud, there is a myth that corruption cannot be measured. This is incorrect; corruption can be and is being measured through a wide variety of innovative approaches (Kaufman, Kraay and Mastruzzi, 2006, p. 5). There are two typologies of corruption measures that this evaluation will focus on, these being qualitative and quantitative by nature. The review will commence by exploring perception surveys before moving on to consider some of the quantitative measures. For the purpose of this evaluation, corruption is defined as ‘the misuse of power in the interest of illicit gain’ (Anderson and Heywood, 2009, p. 748).

Transparency International’s Corruption Perceptions Index was first produced in 1995 and draws upon multiple data sources to calculate the extent of corruption by country, currently 117 countries and territories. In fact 13 data sources were used to construct the 2013 index, including World Bank, World Economic Forum, Economist Intelligence Unit and Political and Economic Risk Consultancy. These datasets were collected in differing years ranging from 2011 to 2014. This survey only measures perceptions, and there is no standard definition of corruption as each of the incorporated surveys operates with its own understanding. Each survey may focus on different aspects of corruption such as bribery of public officials or embezzlement for example. No statistical data such as prosecutions, reported cases or proven incidents of corruption are included. Furthermore, historically the number of participating countries has varied, and thus a country’s position in the table could be influenced by how many countries are covered in any one year. However, since the 2012 index, a revised scoring system has been introduced based on a scale of 0–100 which will be used year on yea...

Table of contents

• Cover
• Title
• 1  The Issue
• 2  Options for Change
• 3  The Dark Figure of Fraud
• 4  Measuring the Cost of Fraud
• 5  Legislating Fraud Loss Measurement
• 6  The Doctrine of Fraud Loss Measurement
• 7  Conclusion
• References
• Index