«Edited by ANNE MASON Research Fellow, Centre for Health Economics University of York and ADRIAN TOWSE Director, Ofﬁce of Health Economics Radcliffe ...»
REFERENCES Broome J (1978) Trying to value a life Journal of Public Economics 9: 91–100.
Broome J (1979) Trying to value a life; a reply. Journal of Public Economics 12: 259– 62.
Claxton K (1999) The irrelevance of inference: a decision-making approach to the
stochastic evaluation of health care technologies. Journal of Health Economics 3:
Dasgupta AK and Pearce DW (1978) Cost–beneﬁt Analysis: theory and practice. London:
Jones-Lee MW (1979) Trying to value a life – why Broome does not sweep clean.
Journal of Public Economics 12: 249–56.
Macran S and Kind P (2001) ‘Death’ and the valuation of health-related quality of life. Medical Care 39: 217–22.
Mishan EJ (1981) The value of trying to value a life. Journal of Public Economics 15:
Mishan EJ (1982) Cost–beneﬁt Analysis. London: Allen and Unwin.
Robinson J (1962) Economic Philosophy. London: Penguin.
Stolk EA, Van Busschbach J, Caffa M, Meuleman EJH and Rutten FFH (2000) Cost utility analysis of sildenaﬁl compared with papaverine-phentolamine injections, British Medical Journal 320:1165–8.
Sudgen R and Williams A (1978) The Principles of Practical Cost–beneﬁt Analysis, Oxford: Oxford University Press.
Williams A (1979) Note on trying to value a life. Journal of Public Economics 12:
Williams A (1985) Economics of coronary artery bypass grafting. British Medical Journal 291: 326–9.
CHAPTER 12 The measurement and valuation of public safety... Paul Dolan and Aki Tsuchiya
INTRODUCTIONPolicy makers strive to allocate limited public resources to where they will do the most good; that is, where the use of resources strikes the best balance between efﬁciency and equity. Determining an efﬁcient allocation may be informed by the results of economic appraisals, which measure and quantify the costs and beneﬁts of alternative allocation decisions. The UK Treasury Green Book (HM Treasury, 2003) recommends that, where possible, costs and beneﬁts are expressed in monetary terms. The notable exception is health where beneﬁts are expressed in terms of quality-adjusted life years (QALYs) (owing much to the work of Alan Williams, of course). This paper considers how the intangible losses from crime and the fear of crime can be measured and valued in ways that allow for the economic appraisal of interventions that seek to reduce crime and/or its impact. We consider how monetary valuations can be obtained and the possibility of developing what Alan referred to as a ‘SALY’ – a security-adjusted life year.
In fact, Alan ﬁrst thought about crime in the late 1960s when he was seconded to the Treasury, and then to the Home Ofﬁce, where he met Vincent Watts. Vincent was working on the ideas of US criminologists Thorsten Sellin and Marvin Wolfgang on quantifying the seriousness of delinquent behaviour, based on how members of the public perceived it. Shortly after, Alan was trying to work out a method of quantifying the seriousness of ill-health, so he contacted Vincent Watts to see if the criminologists’ methodology could be applied in the area of health. This led Alan to Vincent’s wife, Rachel Rosser, who, together with Paul Kind, was doing exactly that. So, the ideas from criminology were applied to health and we are applying the developments in the valuation of health back to crime (see Williams (2005) for this historical circle).
133134 THE IDEAS AND INFLUENCE OF ALAN WILLIAMS
There is now interest in the Home Ofﬁce in valuing the intangible losses from crime, i.e. the difﬁcult-to-quantify losses arising from the emotional and physical effects of crime. The most recent version of the economic and social costs of crime published by the Home Ofﬁce (Dubourg et al., 2005) includes some of the direct costs of crime (such as direct costs to the criminal justice system, the treatment of health losses and indirect costs due to productivity losses), as well as estimates of the intangible victim costs of crime and the fear of crime based upon values we estimated using a QALY-type approach (Dolan et al., 2005; Dolan and Peasgood, 2006). However, those estimates were based on poor-quality data and a number of rather heroic assumptions, and this paper considers ways in which more robust estimates of the losses in wellbeing from actual and anticipated criminal victimisation could be generated.
As with any valuation exercise, there are questions about what is to be valued and how it is to be valued. The simplest thing to value would be categories of crime (robbery, burglary etc.), without describing or valuing the speciﬁc consequences for well-being of those crimes. This has the advantage that data on categories of crime are routinely collected by the police and in the British Crime Survey (BCS). However, these categories are very broad and there is no such thing as a ‘typical’ burglary, for example (Semmens, 2004). In addition, it is virtually impossible to attribute particular fears to particular crimes.
Furthermore, as in the valuation of health, naming the label associated with the medical or criminal cause (e.g. ‘cancer’ or ‘robbery’) will allow respondents to bring all sorts of theories and imaginations of their own to the evaluation exercise that are beyond the control of researchers, starting from what they think causes these problems (e.g. ‘smoking’ or ‘provocation’) to what happens to victims (e.g. ‘side-effects of chemotherapy’ or ‘severe depression’).
Alternatively, we can think of crime and the fear of crime as impacting upon important (and comparable) attributes of our well-being. This is the approach adopted in this paper. The advantage of viewing the effects of crime and the fear of crime in this way means that it will be possible to compare losses resulting from actual victimisation with losses that occur from the anticipation of victimisation. However, as we shall see in the next section, we know surprisingly little about the losses in well-being that result from criminal victimisation, as very few studies have systematically traced the experiences of the victims of crime, and we know even less about how the fear of crime impacts upon well-being. Given the paucity of any reliable data in relation to the type and extent of the losses in well-being arising from criminal victimisation and fear, there is an urgent need for studies of the long-term consequences of crime.
In order to relate ‘epidemiological’ data to valuation data, it is desirable to generate a classiﬁcation system that allows the different attributes of wellbeing that are affected by crime to be combined into overall ‘crime states’, and we discuss what such a classiﬁcation system might look like in the next section of this chapter. This, of course, has many parallels with the work into
THE MEASUREMENT AND VALUATION OF PUBLIC SAFETY 135‘health states’ in the last 20 years, which has seen the development of a number of generic health-state classiﬁcation systems designed to allow the value of each state to be expressed on a single index scale. It could be that we should simply use one of the existing health measures (e.g. the EQ-5D) in a crime context, which would have the great advantage of allowing for cross-sectoral comparisons.
Unfortunately, as we shall also see in the next section, there are important differences between well-being losses from crime and those from ill-health.
Moreover, the losses in well-being from a crime are caused by the wilful intent of the perpetrator (unlike accidents, where injuries may still be caused by others, but not intentionally). This intent to cause harm can result in the victim experiencing losses in well-being when there are no obvious health effects.
Such considerations might result in there being a ‘crime premium’ associated with criminal victimisation as compared to injuries and psychological trauma in other contexts (Dolan et al., 2005).
In the third section of this chapter, where much of the material is taken from Dolan et al. (2007) and where we are grateful to our co-authors Ann Netten and Joanna Shapland, we consider different ways of generating valuations for these crime states. We focus on methods that are preference based; that is, the well-being associated with different states of world is inferred from people’s preferences over those states. As an alternative to preference-based methods, economists have begun valuing non-market goods by considering the effects on an individual’s subjective well-being (SWB) of income and a non-market good and then estimating the required income compensation that would hold SWB constant following a change in the non-market good (Clark and Oswald, 2002). We do not consider this method further here (see Dolan and Peasgood (2006) for a critique of this method and a comparison with WTP).
Economists would typically prefer to infer monetary values from observing consumer behaviour (Atkinson et al., 2004). If we had information about the demand for goods that are intended to reduce the likelihood of being a victim, it might be theoretically possible to tease out the component attributable to preventing the intangible consequences. However, even if we had good data on, say, the demand for burglar alarms, it would be a daunting task to break this down into its component parts. Another possibility is to try to see how the price of accommodation varies between different neighbourhoods with different characteristics, but there are so many factors that may affect the way that property prices vary from one neighbourhood to another that it is difﬁcult to make accurate and robust attributions to any one of them, such as the number of crimes of a particular type. As a result, there is a shortage of useful revealed preference data about the values of preventing the intangible consequences of crime (Lynch and Rasmussen, 2001).
As a result, ‘stated preference’ methods have been developed, which elicit monetary values through hypothetical choices presented to respondents. In this context, this would involve asking respondents to state their maximum
136 THE IDEAS AND INFLUENCE OF ALAN WILLIAMSwillingness to pay (WTP) for a change in well-being from one crime state to another. We discuss the possibility of eliciting values using WTP but, given some serious problems with the methodology, there are doubts about its suitability. We go on to consider methods developed by health economists to allow the calculation of QALYs, and suggest that data generated by a ranking exercise might represent a better way forward.
DESCRIBING THE INTANGIBLE LOSSES FROM CRIME AND THE FEAR OFCRIME The best large-scale evidence on criminal victimisation in the UK comes from data collected as part of the BCS, which now consists of over 50 000 interviews with adults in the UK every year. The BCS reports physical injuries resulting from violent crime such as scratches, cuts, broken nose, chipped teeth etc. However, it does not indicate whether or not the injury received medical attention from a doctor, so the seriousness of the injury is not easily determined. The BCS also investigates the emotional impact of crime by asking questions about which type of emotional reactions were experienced as a result of being victimised. However, it is difﬁcult to make comparisons across waves where the list of possible reactions changes. In addition, the BCS gives no indication of the frequency or intensity of each emotional reaction, and does not even ask how long ago the incident took place, which could be anything from a day to a year ago. What is needed, of course, is longitudinal studies.
Most of the longitudinal studies of victims of crime took place in the 1970s and 1980s and did not use general population samples (e.g. Shapland et al., 1985). For these reasons, it is doubtful whether the results are sufﬁciently generalisable. Research by Denkers and Winkel (1997) is the only recent general population longitudinal study in Europe. They found that victims of crime systematically reported lower levels of well-being than non-victims (less satisﬁed with life, less positive affect, perceiving the world as less benevolent and themselves as less worthy) and, to some extent, higher levels of feeling vulnerable to victimisation (being afraid of crime, people or situations, crime having a greater potential negative impact). However, victims were also less happy than non-victims before the offence (suggesting that those who are less happy have a higher risk of victimisation), so not all the resulting greater unhappiness can be attributed to being a victim of crime.
The intangible losses from the fear of crime are also difﬁcult to deﬁne, not least because the fear of crime is itself a nebulous concept (Hale, 1995). A general concern with fear-of-crime surveys is that they pick up a whole host of things, including emotions that are quite distinct from risk and fear, such as anger (Ditton et al., 1999), and fear and anxieties caused by non-crime activities which people are unhappy about in their environment (Bannister and Fyfe, 2001). In this paper, we deﬁne the ‘fear of crime’ as all the intangible losses in anticipation of possible victimisation (Dolan and Peasgood, 2006).