Psychosocial factors and intergenerational transmission of welfare dependency: a review of the literature
Robyn Penman
Australian Institute of Family Studies
- 1. Summary
- 2. Preamble
- 3. The idea of welfare dependency: a contested notion
- 4. What might culture have to do with it?
- 5. If not a culture of dependency, then what?
- 6. Summing up
- References
1. Summary
This paper explores possible relationships between social, psychological, interpersonal and family factors and the intergenerational transmission of welfare dependency. Three key questions are addressed. First, how does the problem of welfare dependency present itself? Second, what is the evidence for transmission of welfare dependency across generations? Third, what is the evidence for a dependency culture or, if there is not a culture of dependency, what other accounts can be given for some children of parents on welfare also becoming welfare recipients themselves?
From a review of the recent literature, it is concluded that there has been a tendency to treat the problem of welfare dependency and intergenerational transmission as simple matters of fact when the evidence suggests both notions are far more complex. The evidence from the literature does not support the notion of a single causal intergenerational transmission mechanism nor does it support the proposition of a simple culture of dependency dynamic that can generally account for children from income-supported homes becoming income supported later in their lives.
Instead, the literature suggests that a broader, more ecological, or multi-causal, framework with an emphasis on the environmental resources available and capacities of adults and children to use them may be far more useful in understanding how the need for income support can continue from one generation to the next.
2. Preamble
The major thrust of this literature review was to consider the evidence for the role of psychosocial factors in intergenerational welfare dependency. The review began by asking the question: what, if any, psychosocial factors are involved with intergenerational transmission of welfare dependency? However, as the literature search proceeded, it became increasingly apparent that all was not as straightforward as it would seem. The whole notion of intergenerational transmission of welfare dependency was much more complex than is often recognised, at least explicitly, in the literature. The literature was also replete with divergent views and confusing findings.
Because of the complexity of, and confusion in, the literature reviewed, the original question of 'what, if any, psychosocial factors are involved with intergenerational transmission of welfare dependency?' was initially put aside to focus on the evidence for the existence of intergenerational transmission of welfare dependency itself. In addition, a much closer analysis of the way the primary concept of welfare dependency had been, and is being, used was undertaken. In other words, the givens became problematic and the review questions turned to: what is welfare dependency?; what is 'intergenerational transmission?'; and is there any empirical base to the concepts?
The literature on welfare dependency, whatever it may be, is littered with polemic and presumptions. No doubt this is due, in part at least, to the political context in which the issue of welfare dependency is located. The many different researchers and the varying political agendas have meant a range of different definitions of the problem and a range of different proposed solutions. The divergence of views, and sometimes contradictory findings, have been noted by previous reviewers in Australia (for example, Kelleher & Jean 1999; Saunders 2000). However, it was Bonell's (2004) review of 20 years of overseas research on teenage pregnancy-an equivalently political and problem-laden phenomenon-that reinforced the need to approach the review here with great care.
Bonell (2004) noted that there were many different ways in which the fact of teenage pregnancy was defined as a problem and as many different solutions implied. For example, teenage pregnancy has been defined as a problem because of health consequences, because of poverty consequences, because of fertility rate consequences, and because of welfare consequences, among other things. Bonell (2004) also found that there were significant differences between the problem definitions and research conducted in the United Kingdom (UK) and the United States (US). For example, according to Bonell (2004), many researchers from the United States, but none from the United Kingdom, conceptualised teenage pregnancy and parenting as a problem in large part because of its implications for welfare expenditure. In contrast, quite a few UK studies looked at economic factors influencing teen pregnancies, while no research from the United States did so (Bonell 2004). In both of the examples cited above, the differences in the research in the two countries were statistically significant. Bonell (2004) attributes these differences to a combination of political, religious and research design factors.
Bonell's (2004) review acts as a cautionary note that the research literature on welfare dependency will also reflect political, and perhaps religious, factors and will differ substantially across cultures. Feather (2002) has made a particular point about the importance of culture in this area. He emphasises that attitudes to work and non-work are influenced by cultural values and that different social welfare systems, arising from different cultures, have different effects. Cobb-Clark and Gorgens (2004) also point to this difference. Therefore, in reviewing the literature here, priority was given to Australian research. Where overseas research was used, the country of origin is also noted.
Emphasis was also placed on reviewing recent research literature. The literature search was conducted on publications over the past decade, using social, psychological, interpersonal and family as the key search terms. A previous literature review on Transgenerational income support dependence (Kelleher & Jean 1999) conducted by the Australian Institute of Family Studies was used as an additional source, allowing a wider scope for the review than originally anticipated. So, in the end, more literature was reviewed and more questions asked. Following Bonell (2004), this paper starts by asking: how does the problem present itself? The paper then moves on to question the evidence for intergenerational transmission and to consider the role of culture. Finding no evidence for a dependency culture hypothesis, other psychosocial factors that may play a part are then considered.
3. The idea of welfare dependency: a contested notion
How does the problem present itself?
In the introduction to a major collection of articles on Australian welfare, Saunders (2000) describes Senator Jocelyn Newman's intention to reform the welfare system because the rising rates of welfare dependency are placing an increasing burden on government expenditure. This concern is not limited to Australia. Saunders further notes in a later article:
Most governments in the western world have become concerned in recent years about the increasing burden of welfare spending consequent upon the long-term rise in the number of people claiming benefits. (Saunders 2001, p. 29)
The problem of welfare dependency, and the very use of the term, appears to have entered the discourse because of the increase in the amount of money spent on welfare by governments. Bonell's (2004) analysis found this same motivation underlying significant numbers of US research studies on teen pregnancy-teen pregnancy was defined as a problem because of the increasing number of teen parents on welfare.
Curiously, however, figures vary somewhat when presented in support of the observation of welfare increase in Australia. For example, Henman and Perry (2002) state that the proportion of people of workforce age receiving social security payments has grown from 4 per cent in 1966 to 12 per cent in 1980 to 21 per cent in 2000 (roughly one in five). Saunders (2000) states that one in seven Australians of workforce age today relies almost entirely on income support payments, and this represents a much higher level of dependency than in the 1960s and 1970s. On the other hand, Birrell, Maher and Rapson (1997) state that in mid-1995, some 32 per cent of the Australian population aged 20 and over were dependent on welfare, including aged and veterans' pensions. While there can be no doubt that the number receiving income support payments has increased over the past few decades, other aspects of the welfare figures should be questioned.
First, it is not clear in many of the articles reviewed what exactly they meant by being on welfare or receiving income support, and what exactly they meant by being dependent on welfare-however defined. Assuming that this dependency is an economic one, studies using it in this way do not give any detail on how much income support means the recipient is dependent, although it seems that in some instances even a supplemental parenting benefit is counted. In other instances, the condition potentially leading to income support receipt is taken as a proxy measure of welfare dependency (for example, Seth-Purdie 2000 uses unemployment duration). Furthermore, it is often implied that there is an element of pathological, psychological dependency involved. Again, this is rarely discussed explicitly but will be addressed further in this paper.
Bartholomae, Fox and McKenry (2004) point to the problem of measurement in their own work in the United States. Their measurement of being a welfare recipient was a respondent's agreement to having received public assistance welfare. But, as they note, the types of assistance ranged 'from food stamps to energy assistance' (p. 790). Pech and McCoull (2000) in Australia also recognise that there is a need to develop a more sophisticated measure of income support dependence. Certainly in the Australian case there is at least a need to recognise the three different tiers of payment available in the income support system: from pensions (for example, Age Pension and Carer Payment), to allowances (for example, Youth Allowance), to benefits (for example, Family Tax Benefit). Being in long-term receipt of a Family Tax Benefit may be considerably different from being in long-term receipt of a pension for carers.
The key point here is that the concept of welfare dependency, or income support, is generally not well defined and/or is measured in loose and variable ways. Among other things, failure to explicitly define the concepts being measured and/or failure to measure them in the same way creates problems for comparison and for inferences. These variable and sometimes indeterminate measures of welfare dependence used in reports reviewed here can, in part, account for the variable and sometimes divergent findings commented on in the preamble.
Second, it should be noted that the increase in numbers receiving welfare payments in Australia includes a large proportion of those on Age Pensions and Department of Veterans Affairs (DVA) pensions. Although, according to Birrell, Maher and Rapson (1997), even if the aged data is removed there is still a 'very high one-in-five working-age … dependency ratio' (p. 4). Why DVA numbers are still included is unclear. However, what is more important here is that any investigation of factors, psychosocial or otherwise, associated with those on welfare must take into account why they are on welfare. At the least, one would expect that influencing factors for those on a DVA pension would be substantially different from those on an unemployment benefit.
It is particularly important to understand what categories of people are included in welfare data of the sort described earlier, most notably because the issue of transmission of welfare dependency is only relevant to those with young families. Research related to Age and DVA pensions is, in fact, of very peripheral interest to the topic of concern here. So, while there may have been a dramatic increase in the absolute numbers receiving income support payments from the state, the proposal regarding intergenerational transmission and the potential interventionist implications is only applicable to a subset of those on income support: mainly the unemployed and those on disability and parenting benefits.
In considering the make-up of income support recipients the Indigenous population must be kept particularly in mind, given the difficulties under which many of them live their lives (Penman 2004b). Information from Smith (2004) suggests that the changing nature of the relationship between the social security system and Indigenous Australians may need to be taken into account. This relationship can be characterised as a transition from enforced exclusion up to the 1960s, to progressive inclusion over the past 30 years. In 1994, 55 per cent of Indigenous survey respondents stated that some form of social security payment was their main income (compared with 13 per cent of other Australians). This same percentage (55 per cent) was identified in the data of Pech and McCoull (2000) six years later.
While the relative state of Indigenous income dependence has not changed over 30 years (that is, the proportion of Indigenous to non-Indigenous), there has been an absolute increase in the number of Indigenous people on income support. Three major factors have contributed to this increase: a 33 per cent increase in the Indigenous population between 1991 and 1996; a dramatic increase in the Indigenous working age population; and the continued failure of mainstream employment opportunities to keep pace with the needs (and numbers) of Indigenous youth (Smith 2004).
There is no doubt that both the inclusion of Indigenous people in the welfare system and their demographic changes since inclusion would have made a contribution to the burgeoning welfare debt, but this is difficult to estimate. Nevertheless, there may well be a need to consider Indigenous Australians as a special case of income support and this is reviewed in Section 4.
Regardless of how the numbers of income support recipients are derived, the growing concern about the size of the welfare budget brings with it an equivalent growing concern to reform the welfare system (as noted earlier in Senator Newman's remarks). The most concern, and greatest thrust for reform, occurs in Anglo countries where the problem has been construed as one of welfare dependency. The very use of the term welfare dependency reflects a particular set of values and assumptions as well as particular solutions to the problem.
Treating the term benignly, Saunders notes that the 'belief is that dependency on state support corrodes individuals' self-respect and represents a threat to social cohesion' (Saunders 2001, p. 29). With such a presumption it is obvious that the solution is to help make the welfare dependent become independent (of government support).
A number of Australian authors (for example, Birrell, Maher & Rapson 1997; Raper 2000; Henman & Perry 2002) have argued against the very idea of welfare dependency. As Henman and Perry (2002) have put it, neither the concept of welfare dependency nor the similar one of withdrawal from participation has proven adequate in understanding the causes behind the phenomenon of income support increase. The structural realities of labour market and household composition change suggest that cultural welfare dependency or individual failings are not the source of the increase in welfare recipients-certainly not the sole source. Field (2000), from a British stance, argues similarly for the United Kingdom. Further evidence for economic structural changes in the labour market affecting youth in particular is summarised in Kelleher and Jean (1999).
Others working in this field have noted the political motivation underlying the term welfare dependency and have commented accordingly. For example, Strathdee and Hughes (2002) have characterised it thus:
From this psychological/moral perspective [of the government], marginalised youth are deficient, and young people from welfare dependent families have received a cultural inheritance at odds with the Australian government's desire to reduce the number of people on welfare. (p. 37)
Similarly, Travers (1998) describes the concept of welfare dependency as the notion of a cycle whereby dependence, poverty and generally feckless behaviour are perpetuated from one generation to the next. This is the notion of welfare dependency incorporated into the idea of intergenerational transmission that was the initial focus of this review.
From the analysis presented here, it would seem that the notion of welfare dependency is a contested one. But, it must be emphasised, it is not the fact that individuals can be financially dependent on government income support that is being contested; rather, it is the social value implied by the US use of the word dependency. As the authors discussed above implied, calling someone dependent suggests a personal failing, when in fact receiving income support may well arise because of, for example, a labour market failing. In the circumstances, the somewhat more neutral term of income support will be used and caution maintained in considering the idea that dependency of any sort is transmitted across generations.
What is the evidence for transmission?
Pech and McCoull (Pech & McCoull 2000; McCoull & Pech 2000) have undertaken highly relevant research to address the question of intergenerational transmission in Australia. From their review of social security data, they found that most young people from disadvantaged backgrounds do not spend long periods on income support between 16 and 18, although they are more likely to do so than other young people. They also found 'a large proportion of total income support receipt is concentrated among relatively few families, and that there may be little long-term mobility out of the income support system' (Pech & McCoull 2000, p. 50).
On the other hand:
Even among those young people whose parents were apparently the most disadvantaged and income support dependent, only a small minority (about one in six) could be categorised as having been highly income support dependent themselves between the ages of 16 and 18. (Pech & McCoull 2000, p. 63)
In the United States, Rank and Cheng (1995) found that three-quarters of welfare recipients do not grow up in households that received welfare. Therefore, the vast majority of welfare recipients do not come from welfare dependent homes. But, if an individual was raised in a household on assistance, they were more likely to use welfare. These findings from the United States parallel the Australian data of Pech and McCoull (2000) and McCoull and Pech (2000).
So, both the US and Australian research indicates there is an increased probability of children from homes receiving income support becoming income supported themselves. But, both the US and Australian data also indicate that the absolute numbers are comparatively minor-that is, only some go on to be income supported and by no means the majority. Under such circumstances it is unlikely that it is the experience of being in an income-supported family by itself that leads to later income support experiences of the children. If it was the experience of being in an income-supported family that counted, one would expect greater numbers of children becoming income supported later in their life.
On the other hand, neither the Australian nor US studies fully meet the criteria for intergenerational research studies described in Serbin and Karp (2004). First, individuals should be observed at roughly the same age (or developmental stage) in two or more successive generations. Second, the longitudinal information should be prospective rather than retrospective. Third, the data should be multi-level and obtained from multiple measurement sources or domains. Using these criteria, it can only be concluded that there really is no conclusive evidence, one way or the other. McCoull and Pech (2000, p. 110) concluded similarly: 'it would be fair to say that early evidence is mixed' and that 'our early findings must be interpreted with caution' (Pech & McCoull 2000, p. 43).
But, while there does not seem to be any conclusive evidence either way, it must also be borne in mind that the vast majority (five-sixths identified in the early data) of children from incomesupport recipient homes in Australia do not go on to become income-recipient themselves (Pech & McCoull 2000). Similarly, three-quarters of the children in US income-supported households do not go on to become income supported themselves (Rank & Cheng 1995). Put this way, the issue of intergenerational welfare dependency is really an issue of how a small group of people stay on income support from one generation to the next. Put this way, it also becomes apparent that intergenerational welfare dependency is not the sole or even major reason for the massive increase in the numbers of people receiving income support, if, indeed, it is any reason at all.
Given the above data and arguments, the concept of intergenerational transmission of welfare dependency would seem to be problematic. The term welfare dependency itself is a contested notion and to then link it to the idea of intergenerational transmission is to imply a phenomenon of dependency being passed on through the generations. There are also other reasons-considered in the next section-to make one wary of this concept.
What follows is a consideration of the various psychosocial factors that may contribute to one particular pattern of children from income supported homes becoming income recipients themselves, as well as what evidence there may be for why so many do not. In addition, consideration is also given to psychosocial factors that play some part in being a recipient of income support, regardless of parental experience in that regard, and to factors related to poverty in general. At the end of all these considerations the paper returns to the issue of the existence of a transmission mechanism and questions, once again, whether there is evidence for the fact of transmission.
4. What might culture have to do with it?
A dependency culture?
The probability, if nothing else, of children from families claiming income support ending up the same has been observed in various countries. One argument that was advanced by Oscar Lewis in the 1950s to account for the urban poor in Latin America was that there was a 'culture of poverty' (Saunders & Stone 2000). According to Lewis, this distinctive culture arose as a way of coping with the deprivation but the way of coping then went on to perpetuate poverty at the same time.
More recently, this argument has been taken up to account for a similarly observed phenomenon in the United States. Mead (2000) argues that there is a mindset, a set of values, attitudes and beliefs, that he calls 'dutiful but defeated' (p. 48). It is this theory of a particular mindset that is behind the proposal of a dependency culture to account for the appearance of intergenerational transmission of welfare dependency. Indeed, this theory of a dependency culture is critical to the idea of welfare dependency itself.
The theory of a dependency culture arose within an American culture and efforts have been made there to substantiate or disprove it. For example, Rank and Cheng (1995) tested two models: a welfare model consistent with the culture of poverty argument and an economic model. They found that three-quarters of welfare recipients do not grow up in households that receive welfare. Therefore, the vast majority of welfare recipients do not come from welfare dependent homes. They conclude that they found no empirical support for the dependency culture proposition.
Again in the United States, Bartholomae, Fox and McKenry (2004) extended Rank and Cheng's (1995) study and tested cultural and structural models of generational welfare use. They found no evidence for the culture of poverty argument (for example, as described by Mead 2000) but did for the structural model based on parental and public resources accounting for the phenomenon. They concluded that differences in education, socioeconomic status, attitudes and community resources explained the relationship between parental income support and their children's income support-a dependent family culture did not.
Stenberg (2000), in Sweden, tested three types of explanations: cultural–behavioural (for example, culture of poverty), structural–economic (for example, low level of investment in human capital), and policy-related factors (for example, welfare gives poor people incentives to fail). Using Swedish longitudinal data, he found a clear intergenerational effect of welfare dependency that was specifically associated with a combination of social assistance in the family of origin, children's school adjustment, and parental criminality. Children without this combination did not show 'inheritance of welfare dependency'. Stenberg (2000) concluded that, in the Swedish case, there was no single or straightforward explanation for what appears to be intergenerational transmission. Elements of all three types of explanations seemed to be at play. However, Sternberg also recognised that he could not dismiss the possibility of a cultural effect in his findings-that is, peculiar to Sweden. Remarks of Saunders (2001) that 'Australia is not Sweden', need to be borne in mind here.
Saunders and Stone (2000) have undertaken the only direct Australian test of the dependency culture proposition found. They examined whether elements of a dependency culture could be detected among different groups of young people. They found no signs of it in young people between 16 and 18 at school or in young people employed or in training. They did find signs of it amongst some unemployed youth, but by no means all. Moreover, as Saunders and Stone (2000) argued, the weak or broken family ties experienced by the youth in this subgroup makes it difficult to see how a distinctive dependency culture could have been transmitted to them by their parents. Instead, they suggested it was more likely lack of parental guidance towards a work/study pathway that could account for the appearance of a dependency culture amongst some young people. Saunders and Stone conclude from their research that:
If we go looking for a specific dependency culture as the explanation for why parents on income support tend to produce children who go on income support, then we are likely to end up chasing shadows. (Saunders & Stone 2000, p. 132)
In an earlier review of the intergenerational income support literature, Kelleher and Jean (1999, p. 3) concluded:
There is no definitive answer to the question of whether or not there is a culture of intergenerational welfare dependency. If there is a family culture of unemployment or welfare dependency, the literature is divided on how this can be measured, what factors contribute to such a culture and what can be done to prevent the culture continuing into the next generation.
Here, this review can perhaps be a bit more definitive. There appears to be sufficient evidence to suggest that the culture of dependency proposition receives little support. At best, and if anything, the appearance of a dependency culture accounts for only a small group of income support recipients. So what else may be going on?
Rank and Cheng's (1995) arguments may be useful here. They suggest that when income support recipients follow in their families' footsteps it is not because they have learnt the easy life of welfare; rather their parents had financial and economic constraints that limited their investment in the human capital of their children. In support, they refer to Elliot Liebow's study of black street-corner men:
Many similarities between the lower-class Negro father and son do not result from 'cultural transmission' but from the fact that the son goes out and independently experiences the same failures, in the same areas, and for much the same reasons as his father. What appears as a dynamic, self-sustaining cultural process is, in part at least, a relatively simple piece of social machinery which turns out, in rather mechanical fashion, independently produced look-alikes. (Rank & Cheng 1995, p. 8)
The special case of Indigenous income support
There are a number of reasons why Indigenous Australians should be considered as a distinctive group in the considerations here. Not only is their culture and their history different from other Australians but they are, as a group, among the most economically disadvantaged of Australians as well (Penman 2004b). Because of this gross disadvantage, just over half of the population rely on some form of social security payment as their main source of income. This figure does not include the additional numbers of Indigenous people employed in Community Development Employment Projects (CDEP) schemes. On the other hand, while the numbers receiving direct social security support, and 'indirect' CDEP support, are high it is important to bear in mind that the percentage of Indigenous people reliant on welfare has not changed significantly over 30 years (Smith 2004).
Daly and Smith's research in Kuranda and Yuendumu (for example, Daly & Smith 2003; Smith 2004) shows there are many individuals who do cycle through the social security system over their lifetime and that a number of families show the appearance of intergenerational dependency. Moreover, there is concern in Indigenous communities that there are a number of people on welfare who do not know how to work and are not motivated to do so. Pearson (2000) is particularly concerned about the number of Indigenous people reliant on welfare in the Cape York communities and argues that a welfare dependency culture does exist amongst the Indigenous population and that it is poisoning Indigenous society.
Pearson (2000) argues that the 'generally accepted causal chain-racism, dispossession and trauma create social problems which create passive welfare dependency-is wrong' (p. 150). He argues it is wrong on two counts. First, prior to the 1970s, even though racism was rife, and dispossession and trauma facts, the Indigenous population did not have the kind or degree of social problems they do today. Other evidence would corroborate this observation (see, for example, Penman 2004b). Second, Pearson claims that the social problems came after the 'economic condition of passive welfare dependency'-that is, after Indigenous inclusion into the welfare system. There is little doubt that the timing is right, but Pearson provides no other evidence for claiming a passive welfare dependency culture exists. As such, it is not feasible to accept a correlational observation for a causal claim here, especially because of counterarguments by Martin (2001).
Martin (2001) points out, in a constructive consideration of Pearson's argument:
… 'dependency', in terms of a culturally established and validated capacity to demand and receive resources and services … from others, is a core principle through which Aboriginal agency is realised in the structuring of social relationships. (p. 6)
After considering the evidence and the arguments, Martin (2001) concludes:
Pearson is undoubtedly correct in his view that there is evidence of increasing social pathology in the Cape York communities … [but] he is not correct in positing access to welfare incomes for Aboriginal people as contributing to this social breakdown in a direct and causal sense. (p. 11)
From Martin's (2001) point of view, culture does play a role in the appearance of intergenerational welfare dependency, but not quite the one posited in the dependency culture proposition discussed above. Rather than seeing dependency in the pejorative way often implied in the welfare debate (based largely on the moral worth of a productive individual), in Indigenous culture receiving welfare shows a position of strength-that is, they have the power to commandeer from White agencies.
From within Indigenous culture the idea of dependence is quite different. Smith also notes the need to recognise this difference: 'the dependence being experienced by families is quantitatively different from that experienced by other Australian families. It is also qualitatively different' (Smith 2004, p. 56). In particular, it is important within Indigenous communities to speak of the household on income support, not the individuals. It is also important to recognise that most households are not solely dependent on welfare, rather on a mix of social security, CDEP, ABSTUDY, and irregular income from art and crafts.
Even with the risk and resource-pooling that goes on in Indigenous households, these households and the people in them experience substantial economic exclusion from mainstream society (Daly & Smith 2003). This exclusion and its consequences are of such a nature that it is not possible to separate out the profound effect of economic disadvantage and what, if any, cultural dependency may exist. As Daly and Smith (2003) show, not only are there appearances of intergenerational welfare reliance in Indigenous households, but also present are many other factors that are known to correlate with ongoing disadvantage in later life. These factors include the absence of a parent, unemployment status of adults in households, low income, poor educational levels of adults and poor health.
In all, there is no clear evidence that Indigenous households in receipt of income support show a dependency culture based on a set of values that support exploiting the welfare system and avoiding work. While there is evidence that their cultural view can turn receipt of income support from a pejorative to a positive, this is not sufficient to account for the numbers of people in receipt of income support. Many other economic exclusion factors play an equally, if not more, important role in the Indigenous need for income support.
Even though there is no clear evidence for a dependency culture, there is still a strong argument for the need to treat Indigenous people as a separate group in the Intergenerational project. Both their distinctive culture and the depth of social disadvantage warrant it.
5. If not a culture of dependency, then what?
Other specific, psychosocial accounts
The culture of dependency hypothesis has provided a major psychosocial explanatory framework for a pattern of observed correspondence between some parents and then their children on income support. The evidence reviewed here does not support the dependency culture hypothesis for white Australians, though there is some support within the research literature for Indigenous Australians. However, even that support suggests that the notion of dependency within Indigenous culture takes on a different sense than that implicated in the dependency culture hypothesis.
So, if the dependency culture hypothesis does not help explain why some parents on income support have children who later become income support recipients themselves, are there any other psychosocial factors that do? Most of the research literature that addresses this question does not look at intergenerational transmission of welfare dependency as such. Instead, the literature looks at the relationships between psychosocial factors and being impoverished or, more specifically, an income support recipient. And, in most instances, the particular focus has been on unemployment and income support.
Psychological
According to Pech and McCoull (1998), the US research literature tells us that many persistently poor children do not experience poverty in early adult life but growing up poor increases the probability of long-term poverty in adulthood by three times for poor blacks and eight times for poor whites. But:
Researchers have not found that specific psychological and attitudinal factors are important, with the possible exception of attitudes to schooling … This suggests that unfavourable attitudes are not the primary vector for transmission of intergenerational disadvantage. (Pech & McCoull 1998, p. 3)
More recently in Australia, Butterworth (2003) looked at the relationship between mental health and income support and found there is a statistically significant correlation between the incidence of mental health and receipt of income support. Thirty per cent of those receiving income support had a clinically diagnosable mental disorder (substance abuse, anxiety or depressive disorders). Interestingly, it seems that substance abuse is associated more with unemployment and anxiety and depressive disorders than with being a single parent.
Butterworth noted that 'maternal depression and mental illness has negative consequences on a child's development' and then speculated 'it may be, for example, that maternal mental health is an important mediator of intergenerational welfare dependence' (Butterworth 2003, p. 41). The idea of mediating variables is returned to later. But for now, it is important to note that Butterworth's data does not suggest a causal role for mental health, rather just an associative one. If the relationship between type of disorder and type of benefit is considered it is just as easy to argue that being on income support produces the disorder, as the disorder brings about the need for income support.
Social relationships
The concept of social capital has been used broadly to reflect the community resources available to individuals and communities to achieve outcomes. More specifically, social capital has been defined as 'networks of social relations which are characterised by norms of trust and reciprocity' (Stone, Gray & Hughes 2003, p. 3).
Stone, Gray and Hughes (2003) investigated the relationship between social capital and labour market outcomes. They found few significant relationships between social capital and labour force status in the 'informal realm' (that is, family and friends) when they used a core network measure. In contrast, a general measure of social capital type was significantly and strongly related to labour market outcome. Specifically, those with few informal networks, few connections with the wider community and institutions, and low levels of trust were far more likely to be unemployed. But there was no indication of causality or its direction. 'While social capital does relate to labour force status and job search methods used it does so unevenly.' In fact, it looks as if socioeconomic status (SES) is a moderating and influencing variable such that 'the use of family and friends by those from low SES in finding jobs is less likely to result in high quality work' (Stone, Gray & Hughes 2003, p. 23).
Strathdee and Hughes' (2002) qualitative study in New Zealand helps to make more sense of the conclusion of Stone, Gray and Hughes (2003). Basically, the change in the structure of the labour market with less unskilled jobs for young people has meant that while social networks are still being used, they no longer yield high quality employment opportunities. In other words, it is not so much the failure or lack of social capital, but the quality of the employment opportunities available, especially for youth.
These employment opportunities are also affected by where people live. Haveman et al.'s (2004) review of US studies shows that neighbourhood quality and the resources available are related to children's attainments in later life. Birrell, Maher and Rapson's (1997) Australian study shows that there are substantial geographical variations in the percentage of people on welfare support. In particular, it would seem that rural youth are disadvantaged in occupational outcomes, even when parents' own occupation and education is taken into account (Evans & Kelley 2002). Miller (1998) also found that regional characteristics, along with family, were more important in determining unemployment outcomes for youth than personal characteristics.
However, it is not just differences between regions or between rural and metropolitan areas that influence labour outcomes; differences are also noted within metropolitan areas. Kelly and Lewis (1999, 2000) show that there are significant neighbourhood effects of youth employment and unemployment within cities. Youth from neighbourhoods with low SES have higher unemployment. Longitudinal Survey of Australian Youth (LSAY) data also show significant neighbourhood effects on unemployment. Youth from the lowest 20 per cent of neighbourhoods (in terms of SES) experience a higher level of unemployment (Penman 2004a).
These neighbourhood effects could well be contributing to the observed intergenerational reliance on income support. As Kelly and Lewis (1999) note, apparent intergenerational effects could well arise from the geographic location of parents and the increasing financial dependency of youth on those parents. Youth will get less work if they are located in disadvantaged areas and are still financially dependent on their parents. But, most importantly, 'the most substantial determinant of both youth employment–population ratios and youth unemployment is the level of adult unemployment in a neighbourhood' (Kelly & Lewis 1999, p. 14).
Family relations and support
Most of the research literature suggests that single parent families can contribute to poorer child outcomes (for example, Daly & Smith 2003 in Australia; and in the United States, Haveman et al. 2004; Israel & Seeborg 1998; Couch & Lillard 1997). While, in most of these studies, the single parent families were mother-headed, single motherhood does not necessarily lead to poor outcomes. For example, in the United States, Biblarz, Raftery and Bucur (1997) have shown that men raised in a mother-headed family structure do as well as men from two-biological parent families in terms of socioeconomic attainment and social mobility.
Being a single parent, or specifically a single mother, by itself does not appear to be the key influencing factor. Instead, being single and raising children is associated with a constellation of other factors that can play a role in poor outcomes for children. For example, McCoull and Pech (2000) conclude cautiously:
Evidence from overseas, and to a less extent Australia, suggests that family poverty, lone parenthood, parental unemployment and welfare receipt are associated to a greater or lesser extent with less favourable outcomes for some children. (p. 97)
Israel and Seeborg (1998) showed that a number of family variables influenced income performance in the United States, both directly and indirectly. Low-income performance was associated with teenage pregnancy, substance abuse, having a mother-only household and a mother who left school early, low educational expectations, and being black and female. But, most importantly, early dependence on welfare by the respondents' parents was not a significant predictor of the respondents' income levels. According to Israel and Seeborg (1998) 'this set of findings suggests that it is the condition of poverty rather than reliance on welfare that inhibits intergenerational movements out of poverty' (p. 772). In all, these multi-factor studies show that there is a complex clustering of socioeconomic and other family factors at play.
Two Australian studies point to some interesting aspects of these other family factors. Evans and Kelley (2002) found in their multivariate study that 'above and beyond parent's educational experiences, the scholarly resources that parents provide for their children have a demonstrable effect decades later on the children's occupation success' (p. 76). The importance of an early literary culture for children's later education and occupational success has been noted in other arenas (for example, for Indigenous children see Penman 2004b). Indeed, literacy and numeracy achievements in junior high school are the single most important factor influencing employment outcomes into adulthood (Penman 2004a).
McDonald, Brownlee and Greenblat (1993) investigated the possible multivariate effect of locality, parental background and some young person variables on being unemployed and out of school between 15 and 19 years. Only 15 per cent of the variance was explained by these factors and much was due to the young person's age. When age was removed as a variable, only 9 per cent of the variance was explained. What was most interesting however, was that the percentage of the variance explained increased to 21 to 22 per cent when parents' and child's aspirations were included. Lower aspirations were related to outer/rural regions, lower satisfaction with child's educational progress, neither parent completed Year 12 and low income. There is a strong suggestion here that the higher the aspirations of the parents, the more likely their children would continue on at school and get jobs.
There is also an interesting finding from Weatherburn and Lind (1998) that may have implications for the situation of unemployed youth. They show that it is not social disadvantage alone that increases the motivation to engage in criminal acts; rather it is economic and social stress that disrupts the parenting process that leads to a breakdown in restraints on juvenile crime. It makes sense that this same economic and social stress that disrupts the parenting process could also have other negative impacts on youth that in turn lead to disadvantage.
When the above possibility is considered in the light of other studies showing that constructive parenting practices can act as buffers for at-risk children (Serbin & Karp 2004, in the United States) it would seem that parenting styles and the resources available for parenting are an important variable cluster. This is considered in more depth in the next section on developmental accounts.
Developmental accounts: risks and outcomes
From the review so far, it is quite clear that single factors, or even a small cluster of like factors, are inadequate to understand the complex processes at play that lead to some children from income-supported families becoming recipients of income support themselves. Zubrick et al. (2000) recommend that to understand child wellbeing and developmental outcomes there is a need to adopt an ecological framework that respects this complexity of relationships between children, families and communities within a wider social, economic, political and cultural context.
The idea of an ecological framework is a useful tool for identifying sensitivities to the degree of complexity at stake here. Apart from Zubrick et al.'s (2000) proposals, however, no research using the ecological framework to understand pathways to/from being on income support was found. On the other hand, two significant reviews of longitudinal, multivariable studies of child development that show the requisite sensitivity to complexity were found.
The Centre for Community Child Health (2000) in Australia reviewed a number of longitudinal studies looking at risk and protective factors involved in child outcomes. While these studies were primarily from overseas and no specific mention was made of welfare dependency, some general points are still worth noting.
The Centre identified a number of important risk factors that can lead to poor developmental outcomes including, presumably, being in poverty and/or on income support. These factors include: perinatal stress; difficult temperament; poor attachment; harsh parenting, abuse or neglect; parental mental illness or substance abuse; family disharmony, conflict or violence; low socioeconomic status; and poor links with the community (Centre for Community Child Health 2000).
There are clearly a number of psychosocial factors involved here but they cannot necessarily be untangled from within that set of influences or from other more economic or physical wellbeing factor clusters. As the Centre's review shows, risk factors for adverse outcomes often co-occur and they may have cumulative effects over time.
The cumulative effect of familial stressors such as low socio-economic status, young maternal age at birth, large family size and family instability may therefore have a pervasive effect on the well-being of young people. (Centre for Community Child Health 2000, p. 5)
Results from the Australian Temperament Study (Sanson et al. 1991, cited in Centre for Community Child Health 2000) support their claim, as does the more recent work by Seth-Purdie (2000).
According to Seth-Purdie (2000):
[t]he socio-economic gradient across all learning, health and behavioural outcomes appears substantially to reflect differences in the degree to which families are able to meet the developmental needs of young children. (p. 3)
Using longitudinal data, she showed that the more risk factors are present in childhood, the more probable is unemployment, and the longer the duration of unemployment for 16 to 21 year olds in New Zealand. It is of import to note here that it was not any particular combination of risk factors that mattered; rather it was the sheer accumulation of them.
However, it is also important to note from the Centre's review that many high risk children exposed to chronic family adversity (that would include welfare dependency) over long periods of time do not develop intractable problems later in life. Those resilient children had easy temperaments, a good stable caregiver, positive role models/mentors, and positive opportunities at major life transitions (Centre for Community Child Health 2000).
With the focus on dependency, support, poverty or the like, it is easy to forget how many resilient children there are or can be. Focusing on resilience, it is easier to realise that the cited probability of children from income-supported homes going on to become income supported themselves is just that-a probability. And as Seth-Purdie (2000) aptly noted, probabilities are not destinies.
Serbin and Karp's (2004) review of longitudinal studies of intergenerational predictors (but not necessarily causes) of psychosocial risk in the United States takes a slightly different stance, with a focus on early parenthood. They show that early parenthood has been correlated with many forms of future disadvantage in a number of countries, including low parental education, family poverty, single parenthood, low occupation status and job instability (that is, most of the conditions that lead to a need for income support). But, despite this, it is important to note 'the majority of children of adolescent mothers did not go on to become teenage parents' (Serbin & Karp, 2004 p. 340). The minority who did go on to become teenage mothers were more likely to have failed a grade in school, and to be living in poverty and to have experienced negative parenting styles.
There are some important points to draw from Serbin and Karp's (2004) review. First, the intergenerational early motherhood pattern reflects the same pattern found with intergenerational income support-that is, it is a minority pattern. Second, the occurrence of such a minority pattern is related to particular combinations of factors: it is not just the same experience of the parents affecting the children (for example, welfare leading to welfare) but the experience along with other negative/risk factors. Third, parenting style was identified as an important element in this combination.
The potential importance of parenting style and resources was raised earlier and the Centre for Community Child Health's (2000) review also identified this cluster of factors as important for developing resilience (or otherwise) in children. Rickel and Becker (1997) similarly rely on a notion of restrictive or nurturing child rearing practices to account for another intergenerational transmission phenomenon-in this instance child abuse and teen pregnancy.
When the findings about parenting styles are put together with other research, there is the suggestion of a particular dynamic at work here. It is not so much that parents teach (either explicitly or implicitly) their children to rely on income support but that they fail as parents to teach how not to and/or they are unable to provide the basic psychological/educational resources needed to bring about a good outcome. This is the same reasoning used by Saunders and Stone (2000) to account for some youths displaying elements of a dependency culture.
This is not to suggest that those on income support are necessarily inadequate: instead, drawing on Weatherburn and Lind's (1998) argument, it would seem that it is when the economic and social stress experienced by parents on income support becomes great enough, it disrupts their parenting process and this in turn affects their children.
There is some further support for this proposition in the work of Moore et al. (2002) in the United States, which is worth looking at in some detail. They start their work with a very important observation that is very pertinent to any Australian study:
Families experience poverty and welfare in several combinations. For example, families can be consistently poor but go on and off welfare over time; be consistently poor and consistently on welfare; or be consistently poor but never on welfare. However, the relative effects of these patterns on the well-being of children has received little attention. This is a particularly striking omission because there is no typical welfare spell. (Moore et al. 2002, p. 210)
They drew on longitudinal data to assess the effects of different patterns of poverty and welfare on the mathematics achievement, reading skills and social behaviour of 10 to 11 year olds. They found that, after taking background differences into account, children who experienced stable albeit disadvantaged economic conditions did not have worse outcomes than children who were never poor. But children in families whose financial circumstances declined or fluctuated (thus creating stresses without developed coping patterns), were more at risk of behavioural problems and scored lower on literacy/numeracy tests than children who had never been poor. These particular impacts on the children are important because behavioural problems and low literacy levels are two predictors of later unemployment and economic disadvantage.
6. Summing up
A complex issue
There has been a tendency in the literature to treat the ideas of welfare dependency and of intergenerational transmission as simple matters of fact. However, both notions are much more complex than is usually recognised and attempts have been made here to show both the complexities and their ramifications.
As demonstrated in the first part of this review, there is no simple fact of intergenerational transmission of welfare dependency. The very use of the phrase welfare dependency is emotively, if not politically, laden and it has not been readily accepted by all researchers. Indeed, it would be interesting to see what the research might look like if income support, or even poverty, or even more interestingly, income mobility was the predominant focus. All concepts have been used to various effect but it is the idea of welfare dependency here that has proved to be problematic.
The way in which the problem is defined leads to the search for answers in one direction and not another. For example, if the focus was on income mobility then there would be a need to look for change pathways. The question might be what gets people out of poverty, not what makes them dependent on welfare. But here, with a focus on welfare dependency, the mechanism that makes people become dependent, as it were, on welfare has been sought. In particular, the paper was concerned with the specific mechanism that can account for the observation that some children of welfare dependent families also go onto welfare, while others do not.
The mechanism was postulated as an intergenerational transmission mechanism that passed on welfare dependency from one generation to the other. However, no strong evidence for such a mechanism was found. There was some probability evidence and some minority evidence but nothing sufficiently convincing to proceed to identify psychosocial factors affecting the mechanism.
In order to resolve some of the confusion, a distinction is made in this review between addressing the question of intergenerational transmission of welfare dependency and that of psychosocial factors related to economic disadvantage and reliance on income support. Having made that distinction, it has proved of importance in pointing the way to identifying and evaluating two different conceptual frameworks. These are discussed below.
Conceptual frameworks
Transmission mechanism
The idea of poverty and welfare dependency being transmitted from parents to children underlies the search for the intergenerational transmission mechanism. In many ways, it makes sense to conceive of the problem of increasing welfare dependency as one passed on or transmitted from the parents to the children. It makes even more sense when it can be shown that there is a greater probability of children going onto welfare if their parents have also been in receipt of welfare.
Within the understanding offered by this transmission metaphor, the point of research is to look for the conduit or path by which poverty and disadvantage are passed on, as if something is being passed on in the genetic make-up, from parents to children. Such a conception shores up the search for a causal mechanism and for a single pathway (for example, genetic transmission). Yet, from a reading of the literature, there are a number of good reasons why any search for a single, causal transmission mechanism should be treated with caution.
First, for the idea of transmission to be useful, or real, it would be necessary to be able to demonstrate that the causal mechanism lies in something to do with the parents. The evidence reviewed earlier showed that it is not transmitted as part of the family culture. There was no evidence in support of the dependency culture proposition. Indeed, as Saunders and Stone (2000) remarked, if people continued to look for support for a dependency culture, they would end up chasing shadows.
So, if it is not a culture, or world view, that the parents pass on what else could it be? The trouble with this question is that is very difficult (if not impossible) to answer. As Serbin and Karp (2004) argue, not only are there direct biological influences and environmental factors of parenting and socialisation, but there are other things that do not involve a direct 'transfer' of characteristics between generations yet are nevertheless shared. In particular, parents and children have common physical, social and cultural environments (see neighbourhood effects mentioned earlier). These common environments are as likely as anything else to bring about commonalities between parents and their children. Rank and Cheng (1995) argued similarly. They also pointed out that when it comes to impoverished parents there is one further crucial factor: 'The fundamental dynamic underlying intergenerational welfare use is the fact that the playing field is not level as children begin their lives' (Rank & Cheng 1995, p. 681). So in effect, the children of poor or income-supported parents start off their life at a disadvantage and as they develop are exposed to a whole range of factors that they share with their parents, some of which may or may not be passed on.
The second set of considerations has to do with the presumption of homogeneity in the characteristics of those on income support. As Moore et al. (2002) conclude from their US research there is a 'substantial heterogeneity of the population receiving welfare and in poverty. Simple one-size-fits-all approaches are unlikely to be appropriate' (p. 222). So to suggest that there is a single causal mechanism and a single transmission pathway may be to oversimplify the problem. This would seem particularly so when the evidence is considered in terms of relative numbers and not probabilities.
In Australia, Pech and McCoull (2000) found that only one in six children from income-supported homes followed a path onto income support themselves. In the United States, Rank and Cheng (1995) found one in four children fitted this pattern. In both instances, those following the transmission pattern are in the minority. This does not support the hypothesis of a single causal mechanism or pathway. Something more must be going on: something more that means some children become income supported like their parents but most do not.
The work of Stenberg (2000) points out that, at the least, there is a complex constellation of factors with some significant crossover effects. For example, he found that social assistance during childhood lost its intergenerational effect for children without adjustment problems and with non-criminal fathers. On the other hand, he found that the effects of being raised in a lower socioeconomic class and the effects of household income were important among children without adjustment problems and without criminal fathers. But then again, in families that did have such adjustment and criminal problems, the socioeconomic effect was totally eliminated.
Stenberg's (2000) work, more than any other found, shows that there is nothing simple going on here. And, all in all, the literature suggests that the commonality of factors shared by children and their income-supported families cannot be sufficiently or fully accounted for with the idea of a single or discrete transmission mechanism. The issue may well need to be looked at through a different lens.
Making or constructing a life: resources and capacities
Having argued from the evidence that the transmission and related cultural dependency notions are not able to fully or sufficiently account for the observed phenomenon (of some children from income-supported families becoming income support recipients themselves), the question remains about what psychosocial factors may be related to disadvantage/income support.
The evidence reviewed would suggest the need for a more complex multi-fold or multi-causality mechanism to account for the effects of parents on their children's outcomes. Evans and Kelley's (2002) conclusions capture all the findings given in earlier sections and reinforce the need for this more complex and/or multi-causal view. After investigating the multivariate influence of family background and occupation success in Australia they concluded:
All in all, the pattern of family influences on occupational status is a bit like the long-term consequences of a traditional royal christening party in a fairy tale where a wide variety of fairies each presented a different good gift and a few maleficent fairies threw in a curse a two. (Evans & Kelley 2002, p. 78)
There were many different influences identified by Evans and Kelley and most were small-just a nudge here, a nudge there. No one factor held sway and no particular combination of factors necessarily led to substantially better outcomes. But many factors made their contribution.
In the US literature, the resource model of Bartholomae, Fox and McKenry (2004) or the human capital model of Rank and Cheng (1995) go some way towards capturing the greater complexity required. Their approaches place importance on the lack of resources available to impoverished parents, with an emphasis on human capital development of their children. This model may prove more useful than the transmission one and there is certainly empirical evidence in support of it. However, the underlying economic metaphor-based on capital and investment-may prove constraining in the end, especially in terms of acknowledging the role of other non-economic factors.
In particular, neither the resource/human capital model nor Evans and Kelley's (2002) multivariate conclusions capture any sense of agency or active contribution on the part of the child or its parents. This idea of agency has been recognised in the social psychology literature for some decades now (for example, see Harré 1979) and is usually taken to mean that the actions of people are not wholly determined by their environment; rather, they have a say in what they do. On the other hand, this idea of agency does not mean that there is total free will, simply that there is some voluntary base to actions taken.
While this sense of agency is missing from most of the research reviewed-despite it being a common social psychological understanding-there are hints in the literature that it is important. The most noticeable hint is in the descriptions of resilient children in the longitudinal child development literature. For example, consider this description: 'By the time they reached preschool age, resilient children had developed a coping pattern that combined autonomy with help seeking when needed' (Centre for Community Child Health 2000, p. 13). Here the children are coping, being autonomous and seeking help-all active concepts on the part of the child acting as an agent of his/her own life. Yet there is not a single inclusion of these concepts or activities- coping, being autonomous or seeking help-anywhere in the literature reviewed here.
What is suggested here is that the idea of a series of 'nudges', or a series of 'investments', from the child's environment, needs to be combined with the idea of agency on the part of the parents and the children themselves. To do this, a broader, more inclusive or ecological, framework or metaphor is needed. One such framework that suggests itself is one of making or constructing a life. With this framework, the question of how the roughly five out of six children from an impoverished background go on to construct a life that does not rely on income support would need to be asked. What do their parents do that works? How do the children manage? What other resources are the parents, and the children, able to call upon?
The general proposition of Seth-Purdie (2000) is useful as a guide here: it is the degree to which families can meet the developmental needs of young children that matters to how they grow up. In addition, it is the capacity of the children to use what the families have to offer that also matters. There are a number of important economic factors at stake in any parent's capacity to meet their child's developmental needs but these were beyond the scope of consideration in this review. Instead, focus has been on the psychosocial environment only. But it should be noted that this broad framework suggested here can accommodate the psychosocial and the economic.
From the psychosocial literature reviewed here, it would seem that parents are best able to meet the developmental needs of their children if they have access to community resources and social networks; provide a literary culture and other scholarly resources at home; have a nurturing parenting style and create an harmonious home life; and do not suffer from mental health problems or undue stress, especially that arising from fluctuating or decreasing income. But, that is not all. The children also play a role in the process. In disadvantaged circumstances, it is the children with easy temperaments and the capacity to find good role models who will be the more resilient. Furthermore, from the literature reviewed, it may be some complex combination of resources, skills and risks-the nudge here and there-or it may be some particular, but unusual, combinations like that of criminality, social assistance and poor school adjustment as found by Stenberg (2000) in Sweden.
References
Bartholomae, S, Fox, JJ & McKenry, PC 2004, 'The legacy of welfare: economic endowments or cultural characteristics?', Journal of Family Issues, vol. 25, no. 6, pp. 783–810.
Biblarz, TJ, Raftery, AE & Bucur, A 1997, 'Family structure and social mobility', Social Forces, no. 75, cited in J Kelleher & C Jean 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Birrell, B, Maher, C & Rapson, V 1997, 'Welfare dependence in Australia', People and Place, vol. 5, no. 2, pp. 68–77.
Bonell, C 2004, 'Why is teenage pregnancy conceptualized as a social problem? A review of quantitative research from the USA and UK', Culture, Health & Sexuality, vol. 6, no. 3, pp. 255–272.
Butterworth, P 2003, Estimating the prevalence of mental disorders among income support recipients: approach, validity and findings, Policy Research Paper No. 21, Commonwealth Department of Family and Community Services, Canberra.
Centre for Community Child Health 2000, A review of the early childhood literature, Department of Family and Community Services, Canberra.
Cobb-Clark, D & Gorgens, T 2004, The path to self-reliance for young people: a review of the literature, report for Department of Family and Community Services, Canberra.
Couch, KA & Lillard, DR 1997, 'Divorce, educational attainment, and the earnings' mobility of sons', Journal of Family & Economic Issues, vol. 18, no. 3, cited in J Kelleher & C Jean 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Daly, AE & Smith, DE 2003, Reproducing exclusion or inclusion? Implications for the wellbeing of Indigenous Australian children, CAEPR Discussion Paper No. 253, Centre for Aboriginal Economic Policy Research, ANU, Canberra.
Evans, MDR & Kelley, J 2002, 'Family background and occupational success in Australia', Australian Social Monitor, vol. 5, no. 3, pp. 73–78.
Feather, NT 2002, 'Some issues for research on the psychology of work and unemployment: comments on the Australian Psychological Society Discussion Paper', Australian Psychologist, vol. 37, no. 1, pp.10–12.
Field, F 2000, 'Welfare dependency and economic opportunity: a response to Lawrence Mead', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
Harré, R 1979, Social being, Basil Blackwell, Oxford.
Haveman, R, Sandefur, G, Voyer, A & Wolfe, B 2004, 'Trends in children's attainments and their determinants as family income inequality has increased', in KM Neckerman (ed.) Social inequality, Russell Sage Foundation, New York.
Henman, P & Perry, J 2002, 'Understanding the rise in income support recipients', Impact, December, pp. 8–9.
Israel, M & Seeborg, M 1998, 'The impact of youth characteristics and experiences on transitions out of poverty', Journal of Socio-Economics, vol. 27, no. 6, pp. 753–776.
Kelleher, J & Jean, C 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Kelly, R & Lewis, PET 1999, The impact of intergenerational effects and geography on youth employment outcomes, CLMR Discussion Paper 99/5, Centre for Labour Market Research, Murdoch University, Perth.
–––2000, 'Neighbourhoods, families and youth employment outcomes: a study of metropolitan Melbourne', paper presented at Family futures: issues in research and policy, 7th Australian Institute of Family Studies Conference, Sydney, 24–26 July.
McCoull, F & Pech, J 2000, 'Transgenerational income support dependence in Australia: early evidence', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
McDonald, P, Brownlee, H & Greenblat, E 1993, Young people, families and disadvantage, Australian Institute of Family Studies, Melbourne.
Martin, DF 2001, Is welfare dependency 'welfare poison'? An assessment of Noel Pearson's proposals for Aboriginal welfare reform, CAEPR Discussion Paper No. 213, Centre for Aboriginal Economic Policy Research, ANU, Canberra.
Mead, L 2000, 'Welfare reform and the family: lessons from America', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
Miller, PW 1998, 'Youth unemployment: does the family matter?' Journal of Industrial Relations, vol. 40, no. 2, pp. 247–276, cited in J Kelleher & C Jean 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Moore, K, Glei, D, Driscoll, A, Zaslow, M & Redd, Z 2002, 'Poverty and welfare patterns: implications for children', Journal of Social Policy, vol. 31, no. 2, pp. 207–227.
Pearson, N 2000, 'Passive welfare and the destruction of Indigenous society in Australia', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
Pech, J & McCoull, F 1998, 'Intergenerational poverty and welfare dependence: is there an Australian problem?', paper presented at Changing families, challenging future, 6th Australian Institute of Family Studies Conference, Melbourne, 25–27 November.
–––2000, 'Transgenerational welfare dependence: myths and realities', Australian Social Policy, no. 1, pp. 43–67.
Penman, R 2004a, What do we know about the experiences of Australian youth? An easy reference guide to Longitudinal Surveys of Australian Youth research reports, 1996–2003, Australian Council of Educational Research, Melbourne.
–––2004b, The 'growing up' of Indigenous children, Department of Family and Community Services, Canberra. Reprinted in 2006, The 'growing up' of Aboriginal and Torres Strait Islander Children: a literature review, Occasional Paper no. 15, FaCSIA, Canberra.
Rank, MR & Cheng, L 1995, 'Welfare use across generations: how important are the ties that bind?' Journal of Marriage and the Family, vol. 57, pp. 673–683.
Raper, M 2000, 'Examining the assumptions behind the welfare review', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
Rickel, AU & Becker, E 1997, Keeping children from harm's way, American Psychological Association, Washington, DC.
Sanson, A, Oberklaid, F, Pedlow, R & Prior, M 1991, 'Risk indicators: assessment of infancy predictors of pre-school behavioural maladjustment', Journal of Child Psychology and Psychiatry, vol. 32, pp. 609–626, cited in Centre for Community Child Health 2000, A review of the early childhood literature, Department of Family and Community Services, Canberra.
Saunders, P 2000, 'Issues in Australian welfare reform', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
–––2001, 'Australia is not Sweden: the welfare state and national cultures', Policy, vol. 17, no. 3, pp. 29–32.
Saunders, P & Stone, W 2000, 'Australian youth and the dependency culture', in P Saunders (ed.) Reforming the Australian welfare state, Australian Institute of Family Studies, Melbourne.
Serbin, LA & Karp, J 2004, 'The intergenerational transfer of psychosocial risk: mediators of vulnerability and resilience', Annual Review of Psychology, vol. 55, pp. 333–363.
Seth-Purdie, R 2000, 'Accumulated adversity and human capital formation: implications for social policy', paper presented at Family futures: issues in research and policy, 7th Australian Institute of Family Studies conference, Sydney, 24–26 July.
Smith, D 2004, 'Welfare dependence, mutual obligation and the CDEP scheme: lessons from community research and an overseas initiative', in F Morphy & WG Sanders (eds) Indigenous welfare economy and the CDEP scheme, ANU EPress, Canberra.
Stenberg, S 2000, 'Inheritance of welfare recipiency: an intergenerational study of social assistance recipiency in postwar Sweden', Journal of Marriage and the Family, vol. 62, pp. 228–239.
Stone, W, Gray, M & Hughes, J 2003, Social capital at work: how family, friends and civic ties relate to labour market outcomes, AIFS Research Paper No. 31, Australian Institute of Family Studies, Melbourne.
Strathdee, R & Hughes, D 2002, 'Changes in young peoples' social networks and welfare reform in Australia', The Drawing Board: An Australian Review of Public Affairs, vol. 3, no.1, pp. 33–50.
Travers, P 1998 'Welfare dependence, welfare poverty and welfare labels', Social Security Journal, no. 2, pp. 117–128, cited in J Kelleher & C Jean 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Weatherburn, D & Lind, B 1998, Poverty, parenting peers and crime-prone neighbourhoods, Australian Institute of Criminology, Canberra, cited in J Kelleher & C Jean 1999, A literature review: transgenerational income support dependence, Australian Institute of Family Studies, Melbourne.
Zubrick, SR, Williams, AA, Silburn, SR & Vimpani, G 2000, Indicators of social and family functioning, Department of Family and Community Services, Canberra.