Research Paper no. 8 2010–11
Matthew Thomas and Luke Buckmaster
Social Policy Section
15 December 2010
- Governments are increasingly called upon to introduce paternalist policies—that is, policies that restrict the choices of individual citizens in their own interests and without their consent.
- Paternalist policies are often controversial, not least because they infringe a key principle of liberal societies; namely, that citizens are best placed to know their own interests.
- While paternalist policies are often contentious, they are nevertheless ubiquitous. This suggests that the main issue is not whether or not paternalism itself is justifiable, but rather the conditions under which particular paternalist policies may be said to be justifiable.
- This paper argues that paternalist policies may be considered justifiable under circumstances where high stakes decisions are involved, the decisions being made by individuals are irreversible and it is possible to identify failures in people’s reasoning. It is further argued that if paternalist interventions are able to be justified in terms of people’s own values and preferences, then this adds weight to their acceptability given that they do not undermine people’s autonomy.
- Relatively little scholarly attention has been devoted to the questions of what particular forms of paternalism may be deemed to be appropriate. This paper suggests that the principles of discrimination, proportionality, accountability and efficacy provide a framework with which to consider the appropriateness or otherwise of various forms of paternalist intervention.
The authors would like to thank Dr Simon Barraclough and Dr Sarah Miskin; and colleagues, Brenton Holmes, Dr Mark Rodrigues and Michael Klapdor for their insightful comments and assistance in the preparation of this paper.
Governments are increasingly called upon to introduce policies designed to change the behaviour of individuals. This is particularly the case as more becomes known about the cost to individuals and society of certain forms of personal behaviour, such as gambling, consumption of unhealthy food, alcohol abuse and smoking of tobacco. In Australia, this has led to an increased policy focus on areas such as preventive health, gambling regulation and behaviourally based welfare reform.
This latter development reflects a paternalist turn in the welfare policies of many western democracies over the past two decades, known as the ‘new paternalism’. According to this approach, governments are seen as having an active role to play in the re-shaping of citizens’ behaviour—and, in particular, those citizens who are income support recipients and who are especially disadvantaged. Such paternalist policies often involve the close supervision by governments of the lives of poor people who have become dependent on income support. This approach to welfare policy denotes a general shift in emphasis away from simply helping the poor through the provision of income support and various welfare services with relatively few demands made. The new approach attempts to control peoples’ patterns of behaviour ‘in their own interests’.
Many government measures aimed at changing the behaviour of individuals entail some level of coercion. That is, they require governments to restrict the choices of their citizens in certain areas.
The idea that governments may restrict the choices of individual citizens for their own good is commonly known as paternalism. Paternalist policies are often criticised on the grounds that they amount to an infringement of liberty. Further, some argue that paternalism is ultimately counterproductive in that it inevitably leads to dependence on governments and diminishes self-reliance. In some cases, people who criticise paternalism in one area (say, welfare reform) may be supportive of it in another area (say, tobacco control).
Examples of paternalism in everyday life are ubiquitous and often enjoy strong community support: motorcyclists are required to wear helmets, workers are required to contribute to a superannuation fund, parents are required to ensure their children attend school, people may not purchase drugs deemed to be harmful. Nevertheless, the introduction of paternalist policies often provokes heated debate.
This suggests that the central issue is not so much whether paternalism is legitimate, but rather the particular conditions under which specific paternalist policies might be legitimate. This paper examines key issues related to the appropriateness of paternalism in particular policy areas.
The paper begins by defining paternalism and briefly outlining different variants of paternalist policies. The paper then examines the main objections to paternalism, noting that there is a general presumption against paternalist policies in liberal-democratic societies based on the principle that individuals are the best judge of their own interests. Drawing mainly on the work of the political philosopher, Robert Goodin, the paper then goes on to argue that, notwithstanding this presumption, there may be certain instances in which paternalism may be justifiable. Following this, the paper suggests that, while paternalism may be theoretically justifiable in certain circumstances, this still leaves some crucial practical questions to be answered. For example, what limits, if any, should there be on the form of paternalist policies? What kinds of evidence are necessary in deciding whether a paternalist intervention is necessary or has been effective? The paper examines this latter question using as an example the debate over the role and adequacy of evidence for compulsory government management of welfare payments—known as income management. In considering these questions, the paper sketches out a general framework for evaluating the appropriateness or otherwise of paternalism in particular policy areas.
There is substantial debate concerning precisely how paternalism should be defined. Broadly, though, the term paternalism has both a general and a specific meaning.
In general terms, paternalism refers to ‘government as by a benign parent’. That is, the notion that those in positions of power have, just as in the relationship between parents and children, the right and the obligation to overrule the preferences of those deemed incapable of knowing their true interests. Thus, in the area of politics and public policy, paternalism is commonly used in a broad sense to refer to any intervention in private decision-making and/or elitism on the part of governments or other authorities. For example, according to one critic of the former British Labor Government, one of that Government’s least attractive features was ‘its paternalistic contempt for ordinary people’.
In its more specific sense, though, paternalism can be said to have three essential elements, each of which must be in evidence if an act is to be categorised as paternalist. For an act to be said to be paternalist it must: involve interference in a person’s choice or opportunity to choose; be with the objective of furthering the person’s perceived good or welfare; and be made without the consent of the person concerned. This stricter concept of paternalism will be the main focus of this paper.
In relation to public policy, then, paternalism refers to the interference of a government or state with a person without their consent and defended or motivated by a claim that the person interfered with will be better off or protected from harm. Paternalist policies seek to advance people’s (perceived) interests and welfare at some cost to their liberty and freedom of action (autonomy and freedom).
The interference may compel a person to either undertake or refrain from particular activities that affect them. For example, in the first instance, a government or state may require people to save for their future retirement through compulsory superannuation contributions. In the second instance, a government or state may impose a significant rate of excise on tobacco and thereby encourage people to either quit or not take up smoking.
A fourth criterion for categorising an action as paternalist may also be identified. For an action to be paternalist, it must be something more than a government action to correct what economists refer to as market failure in the form of inadequate or imperfect information.
According to an advocate of this approach, public policy researcher, Bill New, an example of this kind of failure would be the case of a person who plans to go swimming in the sea, but who, as a result of their not knowing that the tides in the area are dangerous, may be swept away to their death. In a case such as this, he argues, were the state to intervene through the provision of warnings about the tides or lifeguards to secure people’s safety, then this would not amount to paternalism. According to this approach, state interventions to remedy failures of information such as these are non-paternalistic because ‘the imperfection is not the result of an individual’s inability to best assess their own interests’. Here, the state is intervening in order to achieve an optimal outcome for the individual involved under circumstances in which the market, so to speak, is failing to operate in a manner that enables this person’s best interests to be realised.
If, however, the state were intervening to correct a failure of reasoning on the part of the individual rather than a failure of information, then this would, according to New, constitute a paternalist intervention. To return to the previous example, if the state were to intervene to prevent the swimmer from entering treacherous waters—despite an adequate supply of information (warning signs and lifeguards)—then this would represent a paternalist intervention. This is because the swimmer is being prevented from making their own choice. In the area of drug policy, a similar distinction would be between a government’s provision of information (for example, through advertising) about the harmful effects of certain drugs (thereby correcting market failure) and its limiting access to these drugs on the grounds that individuals were choosing to ignore such information (paternalism).
One difficulty with this analysis is that it is not easy to clearly separate failures of information (market failures) from poor individual reasoning or deliberate disregard. According to economist, Paul Calcott, ‘inadequate information can be a result of the individual’s choices and cognitive and volitional failures can result from institutional failings. An individual may rationally or irrationally choose not to seek further information before making a decision’. As a consequence, Calcott argues that the distinction between failures of information and of individual reasoning cannot be used to differentiate between paternalist and non-paternalist interventions. On such a view, interventions aimed at addressing failures of information may be regarded as paternalist.
Nevertheless, while there are problems associated with precisely distinguishing between information failures and failures of reasoning, we contend that it remains legitimate to attempt to make the distinction where possible. This is because there is something implied in the term paternalism which suggests it should be reserved for those actions that can be regarded as particularly interventionist and beyond the normal course of relations between the state and its citizens. As discussed above, paternalism implies a parental relationship between the state and its citizens (with citizens in the role of the child) whereby the state has a right and duty to overrule the preferences of those who do not know their best interests.
This, we consider, goes beyond actions in which the state responds to market failures of imperfect information through the relatively non-coercive means of information provision. In such cases the state is simply fulfilling its responsibility to ensure that citizens are informed about matters that concern them. While not all citizens may want or need the information, as a matter of practicality (rather than paternalism) the information is provided to the entire group or population, rather than to specific individuals or groups. The provision of such information cannot be said to overrule an individual’s preference in any meaningful sense. Therefore, the definition of paternalism employed in this paper emphasises failures in reasoning over information failures.
A range of different forms of paternalism (themselves existing on a spectrum) are frequently referred to by those attempting to define the concept. Most of these variants of paternalism can be related in one way or another to the definition outlined above. These include:
- ‘soft’ versus ‘hard’
- ‘broad’ versus ‘narrow’
- ‘weak’ versus ‘strong’
- ‘pure’ versus ‘impure’
- ‘moral’ versus ‘welfare’ and
- ‘old’ versus ‘new’.
Soft paternalism refers to the notion that the only conditions under which the state is entitled to intervene in individual decision-making are those in which an individual is considered to be making decisions that are either involuntary or ill-informed. Intervention is only justifiable to the extent that the person needs to be protected from harm to which they did not consent, or to ascertain whether or not they are in fact acting voluntarily and knowledgeably. By contrast, the idea behind hard paternalism is that in some instances it is legitimate for the state to intervene in a person’s actions, even when that person is acting voluntarily and knowledgably. Based on the definition used in this paper, an instance of soft paternalism that was seeking to address possible harm arising from an individual not acting knowledgably should not be considered to be a form of paternalism given that it would be an action correcting market failure.
Broad paternalism refers to any paternalist action, regardless of who performs it, whereas narrow paternalism refers exclusively to state paternalism—that is, the use of legal coercion. An example of the former would be a friend who will not let another friend ride in their car without wearing a seat belt; an example of the latter would be laws against riding in a car without wearing a seat belt.
Weak paternalism is used by some to describe paternalists who consider it legitimate to interfere with the means through which a person is seeking to achieve a particular objective (or preference) if it is believed that these means will not achieve their desired ends (for example, restricted access to a particular medical treatment deemed to be ineffective). Paternalists who believe that it is justifiable to interfere with people’s ends, if such people are deemed to be irrational or mistaken about their ends, are described as strong paternalists (for example, a ban on literature providing information on how to commit suicide).
Impure paternalism describes interventions in which ‘the class of persons being interfered with is larger than the class being protected’. In the case of pure paternalism, the intervention is specifically directed—that is, ‘the class being protected is identical with the class being interfered with’. Philosopher, Gerald Dworkin, uses as an example of impure paternalism attempts by a government to prevent the manufacture of cigarettes on the grounds that they are harmful to consumers—irrespective of the fact that consumers may consent to the harm. This approach would impact on the whole tobacco industry in terms of the manufacture, distribution and sale of the product. Pure paternalism would include measures that try to influence people’s own choice to smoke, such as restricting the locations in which it is legal to smoke, something that would affect smokers but not a broader class of consumers.
Some paternalism theorists also differentiate between moral paternalism and welfare paternalism. Moral paternalism refers to interventions that are calculated to protect people’s moral well-being (or to enforce particular community standards), whereas welfare paternalism is directed towards people’s material or physiological well-being. Dworkin cites, as an example of moral paternalism, the censure of prostitution on the grounds that a person’s selling of sexual services undermines their moral well-being.In contrast, from the position of welfare paternalism, anti-prostitution measures in relation to dangers to physical health (for example, sexually transmitted diseases) would be justified.
As noted above, the last couple of decades have witnessed a paternalist turn in the welfare policies of western democracies, and the emergence of what has been described as the ‘new paternalism’. This form of paternalism marks something of a return to policies of the past which sought to control the lifestyles of the poor. However, new paternalism is different in certain key respects to these earlier forms of paternalism, and it signals a noteworthy change from what may be termed ‘orthodox welfare policy’.
There are two main differences between ‘old’ and ‘new’ forms of paternalism. Firstly, whereas old paternalism was primarily about restricting access to welfare support, new paternalism does not seek to shut people out of aid but, rather, to use aid as a means to an end; that is, to ensure compliance with certain forms and standards of behaviour in order to achieve a designated—often longer-term—benefit or good. Perhaps the most familiar example is that of moving people off welfare and into work for the social benefit of economic participation and the personal good of enhanced self-esteem and autonomy. Secondly, ‘old’ paternalism was typically driven by private organisations, such as churches and philanthropic organisations. These organisations sought to promote education or moral improvement among poor people, usually at a local level and often through closed institutions. By contrast, ‘new’ paternalism is mostly government-led and involves supervision within society.
Old and new paternalist welfare policies can themselves be contrasted with what might be called ‘orthodox’ welfare policies that underpinned welfare state arrangements for the latter part of the 20th century. Generally, under the orthodox approach, welfare was provided on the basis of entitlement (as a right) and recipients were more or less left to choose their own course in life without state interference. Paternalist welfare policies, by contrast, insist that certain behavioural requirements must be met in poor people’s own interests, if they are to receive support. These policies emphasise the conditionality of welfare support.
Unsurprisingly, paternalist policies are controversial. This is largely because they are premised on the notion that the government or the state is better able to make decisions in a person’s interests than the person themself. As Bill New notes, such policies ‘appear to offend a fundamental tenet of liberal societies: namely, that the individual is best placed to know what is in his or her interests’.
This position—that in liberal societies the individual should generally be treated as the best judge of his or her own welfare—is premised on two main grounds, both of which are derived from political philosopher, John Stuart Mill, in his seminal treatise, On Liberty. The first of these grounds is based on the notion that the rightness of an action is determined by its consequences. It holds that because the individual best knows his or her own interests and how to achieve them, the ideal way to maximise individual satisfaction is to not interfere with how people’s choices are made. The second ground insists that allowing people the freedom to exercise choice is not only the best way to maximise utility, but also of intrinsic value; that is, it is good for individuals to be able to exercise choice because this enables them to cultivate and exercise moral autonomy. And, where the state does intervene in people’s lives, taking away their ability to make their own choices, undermines their ability to learn from their mistakes and to develop as morally responsible citizens.
A further, related argument against state paternalism is that there is no guarantee that it will actually improve people’s welfare; indeed, it may make it worse. Thus, for Mill and libertarians following him, ‘the strongest of all arguments against the interference of the public with purely personal conduct, is that when it does interfere, the odds are that it interferes wrongly, and in the wrong place’. This may be because the state is simply not close enough to the individual and their circumstances to understand their best interests. Alternatively, it may be because the state and/or its agents choose to prioritise the interests of others over those of the individual being subjected to paternalist intervention. For example, the state might employ a particular paternalist policy not primarily as a means of genuinely advancing people’s welfare, but rather as a means to realise budget savings.
Another argument levelled against state paternalism is that such interventions can have unintended consequences; in particular, they can help to create what economists refer to as ‘moral hazard’. Moral hazard refers to the idea that an arrangement, which promises people a benefit on the occurrence of certain events, will cause a change in behaviour to make these events more likely. A commonly-cited example of moral hazard is that arising from insurance. It suggests that in situations where a person’s possessions are fully insured, they are less likely to take proper care of them than if they were uninsured. In relation to paternalist interventions, the argument is sometimes made that state interventions to protect people from risk lead them to believe that they are safe and to undertake riskier behaviours as a result. As such, it is argued that paternalist interventions can undermine people’s ability to exercise moral autonomy and their ability to responsibly manage risks.
Typically, then, there is a strong presumption against paternalist policies in liberal societies. This is especially so where the acts being regulated or prohibited through these policies do not cause harm to other people. The rationale behind this precept of liberal societies is clearly outlined in Mill’s harm principle. According to this principle, limiting people’s liberty is only justifiable to prevent harm to other people. On this view, it is not justifiable to constrain people’s liberty in order to prevent self-harm to which the people concerned competently consent. However, according to Mill, where people are considered to be incompetent to make decisions in their own interests, the harm principle permits paternalist intervention. For strict libertarians, the harm principle is the only legitimate grounds for paternalistic intervention by the state.
Perhaps the most obvious recent Australian example of a paternalist policy that could be said to have invoked the harm principle is the policy of income management. Under this policy, a percentage of the income support and family payments of certain people is set aside to be spent on priority goods and services, such as food, housing, clothing, education and health care. Those subject to income management include certain categories of welfare recipient in the Northern Territory (NT) deemed to be ‘individuals at risk’, welfare recipients in the NT and Western Australia (WA) whom a child protection officer has referred to Centrelink to have their income managed, and welfare recipients in Cape York whom a statutory body, the Family Responsibilities Commission, has ordered should be subject to income management for engaging in dysfunctional behaviour.
A key objective of income management is to ensure that ‘money is spent on life’s essentials and in the best interests of children’. Indeed, one of the key pieces of evidence cited by the current federal Labor Government for continuing the policy of income management in the NT is that ‘more than half of parents interviewed in four prescribed areas for an evaluation of income management compiled by the Australian Institute of Health and Welfare (AIHW) reported that their children were eating more and were healthier’. Thus, under income management, certain parents in receipt of income support are having their liberty to spend their allowance as they see fit curtailed in order to ensure that they spend this money on the needs of their children, rather than on, for example, alcohol and gambling.. In short, the policy seeks not only to further the perceived interests of affected income support recipients themselves, but also to prevent harm to their unconsenting children (as well as to the general community).
Thus, in reference to the plight of children, Indigenous leader, Noel Pearson has suggested that critics of income management should:
Ask the terrified kid huddling in the corner, when there’s a binge drinking party going on down the hall, ask them if they want a bit of paternalism … Ask them if they want a bit of intervention, because these people who continue to bleat without looking at the facts, without facing up to the terrible things that are going on in our remote communities, these people are prescribing no intervention, they are prescribing a perpetual hell for our children.
However, use of the harm principle carries with it a number of problems and areas of disagreement. For the purposes of this paper perhaps the key area of disagreement relates to the question of which acts can be said to harm only the person whose behaviour is in question.
It is difficult, if not impossible, to identify an act that harms an individual that does not also harm others, even if this is only indirectly and remotely. For example, a person may choose to ride their motorcycle without a helmet and suffer traumatic head injury as a consequence. While this person may very well have consented to the risks of riding without a helmet, and thought that the act was private and self-regarding, many people who did not consent are likely to be harmed by the act. This could range from the person’s emotional and financial dependants to taxpayers who support the ambulance services and public hospitals on which the person is likely to be reliant. Demanding that motorcycle riders wear helmets in the interests of such others could be seen as an instance of what is termed the ‘public charge’ argument for state intervention: ‘it is not paternalistic, since it is directed against harm to unconsenting others, not against self-harm’.
But if the public charge argument can be invoked in the case of motorcycle helmet laws because of the potential harm posed to unconsenting others then, arguably, a whole range of state interventions may be considered to be justifiable, and not paternalistic. As philosopher, Peter Suber argues:
We can prohibit eating fatty foods on the same grounds. In a welfare state which shifts costs to compensate those who harm themselves, virtually all self-harm will be other-harm too; hence, virtually every corner of life could be regulated by law without violating the harm principle, and virtually all paternalism would be justified.
The main point made by Suber is that a given public policy may or may not be paternalistic, depending on the rationale behind it and the way in which harm and consent are defined. In the first instance, ‘if [the state] believes the act is self-regarding, then it is being paternalistic; if it accepts the public charge argument, then it avoids paternalism and acts under the harm principle’. If the state wishes to avoid the charge of paternalism in relation to a particular policy intervention, then it can do so by widening the definition of harm. For example, if it is argued by the state that smoking in public areas is harmful to others, then prohibiting this act is justified under the harm principle and not paternalistic. The state can also avoid accusations of paternalism by narrowing the definition of consent. For example, if it is determined that disadvantaged people who gamble their limited income support payments on poker machines are effectively being manipulated into doing so, then it may be argued that their consent is invalid. On this understanding, state interventions to restrict these people’s ability to gamble using poker machines might not be understood as being paternalistic.
This last issue—what constitutes competent consent and when may the state be considered to be a better judge of people’s interests than these people themselves—is a vexed one. It is also, unsurprisingly, the source of most of the debate over when it is and is not appropriate for the state to intervene in people’s lives through (paternalistic) public policy. Two key contributors to the debate are Goodin and New, both mentioned earlier, and both of whom isolate circumstances in which the paternalist intervention of the state may be justifiable.
While, as outlined above, it is difficult for many people in liberal societies to accept paternalism in principle, New observes that it is ‘equally difficult to avoid in practice’. Indeed, all but the most dogmatic adherents to libertarian or choice theories recognise the inevitability and in some cases, potential benefits of paternalist interventions on the part of the state. However, it is widely accepted in Western democracies that where states and governments intervene in individuals’ autonomy, this must be subject to rigorous justification.
The objective, according to some commentators, should be to strike a balance between interventions aimed at protecting people’s choices (rather than promoting their interests) and those that try to impose on people what is judged to be objectively good for them. Political philosopher, Robert Goodin, has argued that the best way to achieve such a compromise is to search for some justification for paternalist intervention in the person who is being subjected to this intervention. This entails finding in the person’s own values reasons for saying that paternalism is in their interests. Under such an approach, the emphasis would be to focus on identified failures in the person’s reasoning—failures that can be regarded as standing in the way of their identifying and acting so as to realise their best interests.
Thus, Goodin identifies various conditions under which, he argues, paternalism can be said to be justifiable, without undermining the fundamental principle of respecting people’s autonomy. Firstly, he identifies two key threshold characteristics for the justification of paternalist public policies. As Goodin sees it, the state should only intervene in instances where high-stakes decisions are involved and/or where decisions are more or less irreversible. Goodin cites, as an example, the decision to begin taking addictive drugs. This decision is clearly high stakes in that it would be likely to heavily influence a person’s subsequent life prospects. At the same time, these decisions are to a large degree irreversible in that, having started taking addictive drugs, a person is likely to experience great difficulty in overcoming any subsequent addiction.
Secondly, Goodin identifies several requirements that, he argues, must be met if people’s choices are to be not respected, and state interference considered justifiable. As Goodin sees it, paternalism is only justifiable in instances where public officials better respect a person’s own preferences than the person might have done through his or her own actions or choices. This means that public officials should not interfere with people’s choices so long as they are convinced that these people are acting on preferences that Goodin terms ‘relevant preferences’, ‘settled preferences’, ‘preferred preferences’ and their ‘own preferences’.
According to Goodin, where people make factual mistakes in their reasoning (they may, of course, be led to do so by advertisers) and thus their surface preferences (to smoke or gamble despite their limited income) undermine their own deeper preferences (to stay alive, not be ill or in poverty), then it may be justifiable to override their surface preferences in favour of their ‘relevant preferences’.
People’s reasoning may be less than optimal in a number of circumstances. An individual may not have the technical ability to reason or to adequately work through information in some situations. This is especially so where there is a great deal of information to be evaluated and/or the information being considered may be too complex or conflicting for the individual’s cognitive abilities. It is also the case that an individual may have abstract knowledge or information in a particular area, but be lacking in first-hand experience. For example, they may not have experienced the consequences of a decision, like crashing a motorcycle whilst not wearing a helmet.
In instances such as these it may be argued that the state, with its access to technological resources and institutional memory, is likely to be in a better position to make a rational decision on behalf of individuals and in the interests of the collective. However, it can neither be assumed that this is the case, nor that state intervention is necessarily warranted.
As discussed earlier in this paper, where people have too little information, or inadequate information, then the state cannot be said to be paternalist when it intervenes to provide such information. Such interventions may be justified on the grounds that they are correcting a failure of market exchange, in which there is insufficient information for individuals to make a proper judgement in their own interests. An example of such a situation would be one in which the state required poker machine operators to display some indication of the odds of winning and losing on poker machines. Given that poker machines are the only form of commercial gambling game in which the odds of winning or losing cannot be readily calculated by the gambler, and that this mathematical information is not currently provided to consumers, such a requirement could not be said to be paternalistic.
Goodin also argues that where people’s preferences are short-lived or temporary, rather than ‘settled’, this may justify, on moral grounds, paternalist intervention. For example, people (especially young people) may choose to smoke and claim that that they are happy to accept the associated risks. However, given that we may almost certainly assume that these people are likely to subsequently regret their choice to take up smoking, especially when its impacts are felt, there are grounds for justifiable paternalism to protect them from their current selves. 
Economist and prominent welfare policy researcher, Julian Le Grand has elaborated on this problem of variable preferences through reference to theories about people’s current and future selves. Individuals, he maintains, may make poor choices with regard to their own welfare because they are myopic; that is, they are ‘too short-sighted to take proper account of the future’. To a large extent, this is to be expected because a person’s current self is not the same as their future self. A person undergoes changes—physical, psychological and material—as they age. They are not the same person at the age of ten that they are 30 years later. The greater the gap in age between the present and subsequent self, the less the continuity of the self—that is, the greater the difference between the present and future self.
The implications of this in terms of arguments for (or against) paternalist interventions are that the greater the distance between the present and future selves, the less weight should be given to decisions made by the present self. This is because the future self has no say in the decisions made by the present self but is nevertheless affected by them. In taking such decisions, the present self tends to treat their future self as though they were a different person, and not give sufficient weight to their future self’s interests in arriving at decisions. In effect, the present self imposes an externality on the future self (this is sometimes referred to in economics as an ‘internality’).
Like Goodin, Le Grand does not consider that state intervention is necessarily required to correct this problem. However, where there are empirical grounds to justify intervention to balance the interests between present and future selves that are not optimal, Le Grand appears to agree with such interventions in principle.
Problems arise where there is a choice to be made by policy makers between people’s present and likely future choices. Goodin points out that it is not a simple matter to decide which of a person’s choices should be prioritised. For example, it could be argued that paternalist interventions are necessary to ensure that the poor spend what money they have on healthy food, rather than on items that are not in their long-term interest, such as fast food, alcohol and tobacco (this is more or less the logic behind the policy of income management of welfare payments). However, an alternative view would be that the difficult conditions under which the person’s present self must exist implies that strong consideration needs to be given to the preference of the present self, regardless of the possible impact on the future self. George Orwell discusses this view in The Road to Wigan Pier in which he examines the living conditions of the working class in the industrial north of England before World War II:
The ordinary human being would sooner starve than live on brown bread and raw carrots. And the peculiar evil is this, that the less money you have, the less inclined you feel to spend it on wholesome food. A millionaire may enjoy breakfasting off orange juice and Ryvita biscuits; an unemployed man doesn't … When you are unemployed, which is to say when you are underfed, harassed, bored, and miserable, you don't want to eat dull wholesome food. You want something a little bit 'tasty'. There is always some cheaply pleasant thing to tempt you. Let's have three pennorth of chips! Run out and buy us a twopenny ice-cream! Put the kettle on and we'll all have a nice cup of tea! … White bread-and-marg and sugared tea don't nourish you to any extent, but they are nicer (at least most people think so) than brown bread-and-dripping and cold water. Unemployment is an endless misery that has got to be constantly palliated, and especially with tea, the English-man's opium. A cup of tea or even an aspirin is much better as a temporary stimulant than a crust of brown bread.
Nevertheless, where a person is, on all available evidence and accounts, almost guaranteed to subsequently regret their earlier choice, then Goodin views this as being a relatively clear-cut case for paternalist intervention by the state in favour of their future self.
People have ‘preferred preferences’—preferences that, despite their making different, contradictory or opposing choices, they would actually rather prioritise. (Bill New describes this in terms of ‘weakness of the will’, whereby people know what is in their genuine best interests but nevertheless make a choice that will lead to a ‘sub-optimal outcome’.) For example, a smoker may want to quit smoking but experiences great difficulty in doing so (many smokers are in precisely this position). Where public policies help people to realise their preferred preferences (through, for example, making it more expensive or inconvenient for people to smoke), then such policies cannot be said to be paternalistic in a morally unjustifiable sense. As Goodin observes, ‘in helping them to implement their own preferred preferences, we are only respecting people’s own priorities’.
Goodin also emphasises that before respecting people’s choices, it needs to be ascertained whether or not these choices are really their own. For example, advertising is not simply concerned with conveying information, but also with implanting preferences in people via powerful techniques and insidious strategies that do not engage these people’s reflective judgement. As a consequence, it may be argued that these ‘implanted’ preferences should not be accorded the same weight or respect as people’s genuine preferences. Although he raises this view, Goodin is not entirely convinced by it. This is because, despite having been manipulated (by advertisers or whomever), people may still deliberately claim these preferences as their own in arriving at their preferences. Nevertheless, Goodin does argue that there is a strong case for severely restricting the advertising and promotion of products that are addictive and damaging. As he notes, such advertising and promotion is less concerned with improving people’s ability to make an informed choice whether or not to smoke, and more with inducing people to smoke in order for a company to profit from a product that has significant health risks and costs.
Goodin’s overall argument (generally shared by New and Le Grand) in favour of paternalist intervention under certain circumstances is a powerful one because it does not resort to the interests of the state in justifying such intervention, but rather confines itself to the reasoning failures of the individual, as well as having regard for the issue of personal preferences. It thus strikes a balance between those who see rights as protecting people’s choices rather than promoting their interests, and paternalists who are trying to impose on people what is judged to be objectively good for them.
To illustrate Goodin’s argument is not only that paternalist policies in relation to tobacco control are good for smokers themselves, but also, more importantly, that they can be justified ‘even in terms of the preferences of smokers themselves’ (emphasis added). This is because smokers do not want to die of smoking related diseases and, while they might currently believe that the benefits of smoking outweigh the costs, they are unlikely to do so in the future. Further, while most smokers would like to quit, they find it difficult to do so. In instances such as these, where some of a person’s preferences clearly deserve precedence over others, then, as Goodin sees it, there is a clear-cut case for justifiable paternalism. Paternalism may or may not be justifiable on one or more of the above criteria and it remains a matter for debate whether or not a given intervention is justified merely by satisfying one or more of the above criteria. However, Goodin insists that the argument against a given intervention simply on the grounds that it is paternalistic is insufficient. Paternalism may always be defensible, depending on whether or not it is justifiable against criteria such as those discussed above.
The above position is useful not only because it presents arguments which justify paternalism, but also because it accommodates the fact that the formation of preferences or choices can be a far more complex act than is generally supposed. In other words, such an approach challenges the ‘common sense’ (or ‘realist’) understanding that we tend to have of choice and preferences. A preference is not simply something that can be taken for granted as the final word. It is not a ‘black box’, but rather a matter that can be subject to deeper investigation. The above analysis suggests that the assumption that a choice made by an individual at a given time and place should always (indeed, necessarily) prevail does not grasp the complexity of the processes by which preferences are made.
This more complex understanding of choice has been explored by number of sociologists, including (most famously), Pierre Bourdieu, who analysed the ways in which individuals in particular social classes distinguish themselves from others through the consumption choices they make. According to Bourdieu, taste (the ability to make ‘the right choices’) is not, as is generally supposed, a naturally occurring attribute, but rather one that is socially learned. The acquisition of taste, he argues, occurs primarily through one’s family (the primary socialisation process)—that is, though ‘total, early, imperceptible learning, performed within the family from the earliest days of life’.
Similarly, philosopher and sociologist, Renata Selacl has challenged the realist notion of individual preferences by questioning the assumption that the formation of individual preferences is best understood in terms of individuals making rational, objective calculations that maximise the benefits and minimise the costs of any situation (known as the rational choice model of human behaviour). Instead, she argues that, ‘even if people think that they have the necessary information to make the best choice available, their decision will be heavily influenced by external factors, such as other people, or by internal factors, such as their own unconscious desires and wishes’.
While the work of authors such as the above suggests the importance of subjecting the idea of choice to greater scrutiny, this is not a ‘green light’ for paternalism. New, in particular, takes very seriously Mill’s argument that when the state intervenes through paternalist policies, it often does so wrongly and in the wrong place. This position is supported and explained by a range of sources. Political scientist, James Scott, has argued that the frequent failure of paternalist and other government interventions over the last several hundred years can be explained with regard to certain inherent features of the modern state—in particular, the tendency to seek to apply administrative categories to people’s lives that bear no relationship to their lived experience and which frequently ignore diversity and crucial local knowledge.
Sociologist, Jurgen Habermas, has argued that modern society is increasingly characterised by intervention into people’s everyday lives by administrative and economic systems (the welfare state, the market), resulting in what he calls ‘colonisation of the lifeworld’. In this process, people feel their lives to be increasingly subject to a logic of efficiency and external control, thereby eroding possibilities for meaningful social and political participation. Others have pointed out that those charged with developing and implementing government policy may themselves be subject to failures of reasoning or ‘cognitive imperfections’.
Hence, it may be argued that there is a substantial burden of proof on the state to defend paternalist policies. If the state is to justify its paternalist policies, then, it must do more than simply demonstrate that it is capable of knowing the individual’s interests better than him- or her-self. Rather, the state must articulate sound reasons for actually believing that it knows the individual’s interests better than the individual. Importantly, even though a state may be able to demonstrate using the above criteria that a person’s reasoning ability is less than optimal, this does not amount to evidence that the state is better able to promote that individual’s interests.
The previous section outlines the basic philosophical case under which a paternalist intervention might be considered justifiable. However, what it does not do is provide much guidance in deciding on the legitimacy of particular forms of paternalist intervention. While, for example, we might agree that a clear-cut case of myopia (such as the decision to smoke) may make some kind of paternalist intervention theoretically justifiable, this then raises some important practical (or consequentialist) questions. Among the most important of these are questions relating to the legitimacy of particular types of intervention (‘are some forms of paternalism more justifiable than others?’); and questions relating to evidence (‘how do we know the intervention is required or will do any good?’).
Relatively little attention has been paid to the nature or form of paternalist interventions. Clearly, there are more or less intrusive—and more or less justifiable—forms of paternalist intervention. Indeed, the emergence over the past couple of decades of new and more substantial forms of intervention (such as compulsory income management) has thrown into relief the issue of whether or not certain forms of paternalist intervention may be justified. The possibility that paternalism might be found to be justifiable in a given situation, tells us little about whether a particular form of paternalism is justifiable. Clearly, this is a critical issue in evaluating the appropriateness and justifiability of a given instance of paternalism.
Given that there has been little theoretical attention given to the appropriateness of different forms of paternalism, this section will suggest four criteria that might be used to establish a framework for evaluating paternalistic interventions: discrimination; proportionality; accountability; and efficacy. These are drawn from various public debates about paternalism and, most importantly, acknowledge the strong presumption against paternalist policies in liberal societies. This presumption, we suggest, can be regarded as placing limits on the form of particular paternalist interventions.
If we accept the principle that states should only engage in paternalism in limited circumstances (for example, where irreversible or life-changing decisions are involved), it could be argued that particular interventions should be designed in ways that avoid an impact on people other than those deemed to be at risk of harm from a particular behaviour. In short, paternalist interventions should discriminate between those for whom paternalism is deemed necessary and those for whom it is not. As outlined earlier in the paper, this position, whereby the class being protected is identical with the class being interfered with, is known as pure paternalism.
While pure paternalism is consistent with the principle of limiting the extent of state interference in people’s preferences, there are also reasonably compelling arguments for impure paternalism (that is, where the group being protected is larger than the group being interfered with).
There may be situations in which a pure intervention is not likely to be as effective as an impure intervention. In pursuing a pure, rather than impure, approach, the state might avoid interfering with the liberty of those not deemed to require protection, but this could come at the cost of assisting some people for whom that protection would be beneficial. For example, a class of people might be particularly susceptible to harm as a result of consumption (or overconsumption) of a particular product—such as young women and sweet alcoholic drinks (alcopops). Where evidence existed that harm could be prevented more effectively through a general (impure) policy approach, such as an increase in taxation, that affects all consumers than through pure approaches, such as targeted information campaigns, strict adherence to a pure paternalism approach would leave more people at risk of harm than an impure approach.
A further argument against a strictly targeted approach is that, by its very nature, it formally discriminates against certain groups of people. That is, it creates a situation in which some people have choices but others do not. For some paternalists, this is the whole point—some categories of people need protection but others do not. For others, however, strict targeting is an even more egregious violation than a more general approach because it sets aside the rights of a particular group of people to choose for themselves. For those taking this stricter approach, the fact that the targeting happens under the guise of protection does not necessarily make this violation legitimate. In some cases, it is argued that such targeting may increase the possibility of harms to certain people precisely because it involves the setting aside of their rights, thereby creating the risk of increased social stigma and humiliation. This argument is frequently made in relation to compulsory income management and other paternalist measures associated with the Northern Territory Emergency Response (NTER). This indicates that while, on the surface, pure paternalism may be more consistent with the liberal presumption against paternalist policies, there may be instances where this more pure approach is inconsistent with other values and objectives, including the justified one of securing assistance for certain groups of people.
Another criticism of the form of income management used in the NT is that it is not discriminating enough in its application. In this connection, it is worth noting a crucial distinction between the income management arrangements operating under the NTER and those instituted in Cape York Peninsula in Queensland. While income management is imposed relatively indiscriminately on certain categories of welfare recipient within the NT (that is, on a categorical and geographical basis), in Cape York income management is only applied to those income support recipients who fail to meet certain obligations to their children and the community. Where a person does not meet one or more of their obligations, the Family Responsibilities Commission may recommend that the sanction of income management be imposed on them. If the person is able to demonstrate that they have been complying with their obligations, then their right to manage their own payments may be reinstated. In other words, the form of income management employed in Cape York is more clearly targeted at those failing to meet their obligations. In this respect, it could be said that the Cape York approach, while still setting aside the rights of certain people to choose for themselves, provides an example of how paternalism may be designed to restrict its impact to those actually in need of assistance.
What is the appropriate level of interference in a paternalist intervention? How deeply may the state intervene in the choices made by individuals? Again, the principle that paternalism should only be applied in limited circumstances suggests that in liberal societies there should also be limits to the extent of intervention that is justifiable.
One such limitation might be that any intervention is the minimum necessary to achieve the effect of protecting those subject to the policy. In other words, paternalist interventions should be designed in such a way as to have the minimum possible level of interference in people’s lives, whilst remaining a viable policy for bringing about necessary change. A further limitation might be that any paternalist intervention should be proportionate to the problem being addressed by the intervention.
Here is it useful to return to the distinction made between strong and weak paternalism (the former interfering with ends, the latter interfering with means) outlined earlier in the paper. Another useful distinction might be between ‘deep’ and ‘shallow’ interventions—the former referring to quite substantial interference (including violations of civil, political and social rights) and the latter to less substantial interference. According to the principles outlined above, strong and deep paternalism would rarely be justifiable—that is, only as a response to threats of the most irreversible, life-changing kind.
Of course, precise judgements about such matters as whether a particular intervention is the minimum necessary are far from simple. Among other issues, such judgements require engagement with evidence about the nature of particular social problems and the outcomes of particular social policy interventions. Issues of evidence are themselves particularly complex and will be discussed in detail in a later section of the paper.
In liberal societies, it is expected that governments should be transparent in their actions and accountable to citizens. This raises the question of whether or not the design of an intervention should make it obvious that some form of attempted coercion is taking place. Relatedly, is it legitimate for the state to make decisions on behalf of its citizens or influence them to make particular decisions without telling them? Are there some instances in which the benefits are so obvious that the consent of citizens may be assumed (and who precisely should decide when this applies)?
Such questions have been subject to significant debate in recent years. Some public health experts, for example, have begun to argue for preventive health interventions that take choices away from individuals. It is generally accepted that some of the biggest gains in preventive health are to be made in areas where individuals do not exercise individual choice (such as water fluoridation, sanitation and food regulation), or where there are measures in place to ensure the widest possible compliance (such as vaccination).
UK public health expert, Nicholas Wald, in arguing for a stronger approach to intervention in nutrition and other areas of public health, has stated:
Contrary to current perception, the key to effective public health is not individual choice but collective action linked to public trust in its value. Most of the main determinants of health vary little among people in a community. The scope for individuals to choose healthy and safe foods, drinks, transport, or buildings is limited; the similarities in exposure are greater than the potential differences.
On the issue of nutrition, Wald argues:
The chief merit of the increasingly popular convenience foods is their convenience. Individuals have little influence over their composition. Even foods that are described as being healthy can be high in sugar and salt, counterbalancing any benefit from added micronutrients, such as folic acid. But discouraging the use of convenience foods is not practical; we need collective action to reduce the amounts of salt, sugar, and saturated fat in foods, and a sensible policy on portion sizes in restaurants.
Public health arguments such as these highlight the difficult choices governments face (in particular, between paternalism and individual freedom) when seeking to make significant gains in the area of preventive health.
Such issues have also been apparent in the debate about what is known as ‘choice architecture’ in public policy. Choice architecture involves influencing the context in which people choose in order to help them to ‘do the right thing’. It seeks to change the environment so as to influence people’s behaviour through ‘nudging’ rather than coercing. For example, governments may introduce policies that try to de-normalise smoking and make it as difficult as possible for people to smoke. Some would argue that such interventions are not justifiable. This is because rather than enhancing smokers’ or would-be smokers’ autonomy by negating the tobacco industry’s influence and informing people of the dangers of smoking, these interventions undermine their autonomy by doing away with their ability to choose. Further, such interventions could be seen as undemocratic because democratic societies generally allow people to make informed choices.
It might be said, therefore, that ‘silent’ or ‘below the radar’ paternalism offends the liberal principles of transparency and accountability. Such approaches also potentially offend the presumption against paternalism in liberal societies by making the extent of paternalism in society less visible to citizens.
Nevertheless, some argue in favour of such interventions on the grounds of efficiency and practicality. For example, public policy researcher, Steven Kelman, has argued in favour of regulation of food additives on the grounds that it may be more efficient for regulators to make decisions on behalf of consumers than to expend resources on information campaigns that may not even reach all of the intended recipients. Indeed, Kelman does not even regard such cases as paternalism because, he argues, they generally do not involve enough genuine disagreement between the state and citizens. The point is that the term paternalism should be used for those situations in which the state imposes its views even on those who disagree. Calcott suggests that, at worst, such cases may be described as ‘impeded agreement’.
The general principle placing limits on the form of particular paternalist interventions also gives rise to questions about the likely effectiveness of any intervention. If it is agreed that instances of paternalism should be relatively rare, then, arguably, there should be a reasonable prospect that any given instance of paternalism should be likely to be effective both in terms of its immediate results and improving the capacity of individuals to act autonomously over the longer term.
A key issue that might be raised in relation to efficacy is, given the potential (discussed above) for paternalist interventions to fail to adequately account for social diversity, the extent to which local conditions should be considered in the design of the intervention. This raises the issue of consultation. Should those subject to the intervention expect to be consulted in the design and operation of any intervention? Again, the Cape York approach to income management has been suggested as an example of where local conditions have been appropriately taken into account in that this intervention emerged from the Cape York communities themselves. In other words, it was a local initiative, rather than one largely imposed by outsiders. The program was introduced by the Bligh Government in Queensland after consultation with the Commonwealth Government and Noel Pearson’s Cape York Institute. Local elders who make up the Family Responsibilities Commission in each community also have a hand in the income management process and are reported to have assisted local people in avoiding the sanction of welfare quarantining. Thus, the procedural rights of those income support recipients who are affected are more likely to be maintained.
Another important issue in relation to the question of efficacy is whether an intervention might actually make matters worse. For example, it is often argued in relation to compulsory income management that any possible good emerging from the intervention is likely to be cancelled out by other unintended consequences, such as shame, stigma, disempowerment and further dependence on the state. On this latter point, it is worth noting that an evaluation of the child protection income management scheme in WA found that 15 per cent of those subject to the program cited ‘not needing money management skills when having their money income managed’ as a reason for not using financial management services. This suggests that in some cases income management may actually perpetuate reliance on the state.
An example of a proposed policy intervention that attempts to strike a balance between paternalism and individual autonomy and responsibility is provided by the Productivity Commission’s recent recommendation for a poker machine pre-commitment system. Under such a system, people with gambling problems could be required, for example, to set a limit on the amount of money they wish to spend on playing poker machines in a given period (a day, week or month) or set a time limit on their use of poker machines. This would be achieved through the use of a smart card which poker machine users would need to sign up for and on which their pre-commitment details would be recorded. All poker machines would require the insertion of a smart card and would restrict a machine user’s play according to the limits recorded on their card.
One of the most significant benefits of a pre-commitment system is that it could help to ensure that problem gamblers were protected from harm while not undermining their autonomy and ability to exercise personal responsibility. As the Productivity Commission explains:
Pre-commitment involves consumers pre-setting the terms of their future gambling in ways that address the harms—small or large—that they associate with gambling. Since the consumer sets these options [albeit because they are obliged to], pre-commitment is consistent with consumer sovereignty. And, just as is the case for self-exclusion, pre-commitment gives people with control problems a capacity for exercising self-responsibility. It is a regulation that reinforces, rather than erodes, personal responsibility.
A further significant issue in relation to the question of efficacy is that of the role of evidence. Is it reasonable to expect that reasonably conclusive evidence, or where evidence is not available, clearly articulated reasons will be provided by the state for why the particular form of intervention is regarded as likely to be effective?
The above questions can be seen as addressing the kinds of issues that would arise in evaluating whether a particular form of paternalist intervention is justifiable from the point of view of efficacy. Arguably, they could form the basis of a framework for evaluating such interventions. Central to any such framework would be the question of evidence. However, as will be discussed in the next section, issues related to evidence justifying paternalism (and social policy in general) are particularly complex.
The previous section raised the question of what evidence is required in order to assess the efficacy of particular social policy interventions. That is, what constitutes evidence-based policy in relation to paternalist interventions. In some cases, the evidence is relatively clear-cut. For example, there is longstanding and substantial evidence for the overwhelming benefit of tobacco control policies for individuals in terms of improved health outcomes and reduced risk of death. Hence, the area of tobacco control could be said to be one in which it is almost always the case that more good than harm is done in the course of paternalist intervention (at least, insofar as health outcomes are concerned).
However, the evidence of clear-cut benefit for other policy areas in which paternalism might be considered is not as unambiguous as it is for tobacco control. In areas such as these, it is not only necessary to ask whether more good than harm is being done, but also which is the greater good—potentially improving people’s material well-being or respecting their autonomy.
One approach to this problem is to consider evidence according to a hierarchy of different evidence types. Such a hierarchy was recently developed by economist (and now member of the House of Representatives in the Australian Parliament), Andrew Leigh:
1. systematic reviews of multiple randomised trials
2. high-quality randomised trials
3. systematic reviews of natural experiments and before-after studies
4. natural experiments using techniques such as differences-in-differences, regression discontinuity, matching, or multiple regression
5. before-after (pre-post) studies
6. expert opinion and theoretical conjecture.
While there has been significant debate about the content of this and other such hierarchies, it is important to recognise that there are some fundamental problems with applying such an approach across the range of policy areas in which paternalist policies may be applied.
One such problem relates to the relative lack of robust, high quality evidence in social policy areas, compared with that available in the area of health policy. Among other things, this reflects the fact that health research has far greater access to funding and resources than social policy research. As social policy researcher, Ilan Katz, argues, ‘this means that the number of randomised control trials (RCTs) or other rigorous evaluations of social policy issues is vanishingly small, despite the exponential growth in the literature’. As Katz further argues, the absence of such evidence undermines the utility of evidence hierarchies for policy-making in the social policy area.
There is also the more fundamental problem that social policy interventions are frequently far too complex to be easily accommodated within the framework of RCTs. As Katz notes, RCTs are ‘best designed for single interventions which have simple and easily measurable outcomes, such as taking Aspirin’. However, social policy interventions are generally more complex than health interventions such as those providing discrete medicines. For example, social policy interventions generally comprise more complex, interrelated elements, making it particularly difficult to measure outcomes with any precision.
The ‘evidence problems’ described above are well demonstrated by the debate over the role and adequacy of evidence related to the efficacy of the policy of compulsory government management of welfare payments—known as income management. The remainder of this paper examines this in further detail.
As noted above, income management refers to arrangements whereby a percentage of the income support and family payments of certain people is set aside to be spent only on priority goods and services, such as food, housing, clothing, education and health care. While the total amount owed to a person subject to income management is not reduced, that person loses the discretion to spend a percentage of their welfare income on things other than those that the Government deems to be priorities.
Australia is the first country to adopt this particular system for controlling how welfare recipients spend their payments. Currently, a person may be subject to compulsory income management for a number of reasons including:
- they have been in receipt of particular welfare payments for a certain amount of time and hence are regarded as ‘at risk’ (currently in operation only in the NT)
- they are the subject of a report by child protection authorities (NT and WA)
- they have been assessed by social workers as requiring income management due to vulnerability to financial crisis, domestic violence or economic abuse and
- they are a resident of Cape York, Queensland and have been referred for income management due to behaviour deemed to be irresponsible.
People may also volunteer to participate in income management.
Compulsory income management may be said to be a form of paternalism because (a) by definition, people who meet the criteria for participation must participate (though, they may apply for an exemption on certain specified grounds) and (b) it is imposed in what the Government regards as the best interests of those individuals (and their dependants) required to participate. That is, a central policy objective of income management is to ‘tackle the destructive, intergenerational cycle of passive welfare’ by welfare enhancing recipients’ ability to move out of welfare dependence and into economic and social participation.
Since its introduction by the Howard Government in 2007 as part of the NTER, compulsory income management has attracted some public support, but also substantial criticism. Much of the debate surrounding income management has related to the question of evidence. That is, is there a sufficient body of evidence to justify the policy of income management? Alternatively, is there evidence of policy failure or harmful consequences arising from income management?
Much of the controversy over evidence stems from the fact that there are substantial difficulties associated with evaluating the effectiveness of income management. For a start, there are very few studies available that seek to directly evaluate the effectiveness of income management. In part, this is because income management is still new and untried elsewhere.
Further, there is the problem of separating the impact of income management from other policy interventions and factors. For example, in relation to the NTER, it is virtually impossible to isolate the various elements (such as income management, alcohol prohibitions and increased policing) in a way that would make it possible to draw reasonable conclusions about the relationship between particular policy inputs and particular outcomes (such as improved health and education outcomes and reductions in crime)—or about the validity of generalising from these outcomes to other proposed income management schemes elsewhere.
There is also the problem of deciding precisely who or what is to be the subject of the research—that is, what is the ‘unit of analysis’. As Katz has argued in relation to the NTER:
How could one know whether it had been effective? Its consists of a wide range of components [and] [n]ot all communities have received all the interventions. In addition some people have moved from one community to another, while many have left the remote NT altogether. So the first problem, before we even get to outcomes, is that it is not clear what the unit of analysis for an evaluation would be—is it the communities or is it the individual children who the intervention was aimed to ‘save’ from sexual abuse and neglect?
Other methodological issues, some of which have become a source of controversy in the debate about income management, include:
- the difficulty in obtaining (or, in some cases, absence of) a comparison group or baseline data against which to measure the impact of income management
- the limited amount of quantitative data—reflecting to some extent the problems associated with collecting such data from remote NT communities—capable of providing evidence of the impact of income management
- in the absence of quantitative data, the strong reliance of government evaluations on qualitative measures based on the perceptions and views of various stakeholders (clients, community members, store owners, community sector employees, Centrelink staff and government business managers) about the impact of income management
- criticism that those responsible for some reviews of the NTER and/or income management have not been sufficiently independent from the Government and
- criticism of other design aspects of various reviews—for example, in relation to the representativeness of data based on sample size, the nature of questions asked and the quality of information provided to participants.
As such, thus far the evidence provided for or against income management, or any other individual inputs of the NTER, has, at best, only ever partially clarified particular aspects of a complex situation. This highlights the difficulties associated with obtaining clear evidence of the efficacy or otherwise of paternalist interventions.
Arguably, it also draws attention to the importance of context, argument and competing principles in the case of policy interventions where it is difficult to draw clear conclusions from the evidence. In considering the ways in which evidence has been used in relation to the NTER, and in Indigenous affairs more generally, political scientist, Will Sanders has argued that evidence should be seen as a small part of a larger process of argument and persuasion. He maintains that in NTER evaluations:
Evidence did not stand alone as demonstrating what worked. It had to be contextualised, interpreted and inserted at appropriate points in arguments between the competing principles of guardianship and choice and the ideological tendencies towards social liberalism and directiveness.
Sanders’ argument is not one against the use of evidence in policy. Rather, it is a call for a more appropriate balance to be drawn between evidence and the competing principles and ideologies in the service of which the evidence is mobilised. The point is that it is possible to argue for or against income management (and paternalist interventions more generally) on the basis of principles like guardianship, equality and choice, and what is an appropriate balance and synthesis between them, without prioritising evidence that is ambiguous and contested.
Paternalist policies are almost inevitably controversial in liberal societies. Not only do they challenge the fundamental principle that individuals should have sovereignty over themselves when it comes to judging and acting on their own interests, but they also have the potential to do more harm than good.
Given these factors, it is important that paternalist policies should be subjected to rigorous scrutiny where it comes to their legitimacy, form and efficacy. This paper has suggested that paternalist policies are more likely to be legitimate under certain circumstances than others. Where high stakes decisions are involved, the decisions made by individuals are irreversible and it is possible to identify failures in people’s reasoning, then paternalist interventions may be justified. And, if interventions are able to be justified in terms of people’s own value judgements, then they may be deemed reasonable given that they do not undermine people’s autonomy. The paper has also proposed a number of principles that could form a framework for the evaluation of whether or not certain forms of paternalist intervention may be more justifiable than others.
While this paper has sought to provide some guidance as to where and in what forms paternalist policies might be said to be justifiable, it has also highlighted the complexities associated with such interventions and their evaluation. It has thus served to illustrate the need for policy makers to be especially clear about their motivations and objectives where it comes to paternalist interventions.
For copyright reasons some linked items are only available to members of Parliament.