Continuity and change in the Community Development Employment Projects Scheme (CDEP).
Hunter, Boyd ; Gray, Matthew
Introduction
The Community Development Employment Projects (CDEP) scheme has
long been an important feature of the labour market for Indigenous
Australians, particularly in remote and regional areas. The scheme
involves participants working for a notional equivalent of their income
support payment. The design of the scheme changed little from its
creation in 1977 until 2009. CDEP organisations were allocated funding
to pay the wages of CDEP participants at a level similar to income
support payments, supplemented with administrative and capital support.
This funding has been used as a means to provide employment, training
(informal and formal), activity, enterprise support and income support
to Indigenous participants. The scheme has always had a strong community
employment and community development focus and, since the late 1980s,
has progressively acquired a labour market program objective of
increasing the job readiness of participants. This focus especially
emerged during the late 1990s. (1)
CDEP participants qualify for additional income above their income
support entitlement in the form of a CDEP 'participant
supplement'. In 2008 this was $20.80 per fortnight. Historically
the income test applied to CDEP payments has been more generous than the
income test applied to income support payments (such as Parenting
Payment and Newstart) and the rate at which CDEP payments were reduced
as non-CDEP income increases has been lower than was the case for income
support payments. (2)
From its creation in 1977 CDEP expanded to have over 35,000
participants at its peak in 2002-03. Since then the number of CDEP
participants has declined. The scheme was progressively withdrawn from
non-remote areas from 2007, a process which was completed by 2009 (ANAO
2009). Collectively these changes have seen the number of participants
decrease from about 35,200 in 2002-03 to about 10,500 in 2011. (3) From
mid-2013 the new Remote Jobs and Communities Program will bring together
the Community Development Employment Projects, Job Services Australia,
Disability Employment Services and the Indigenous Employment Program
into a single integrated service (Australian Government 2012) and CDEP,
at least as it has historically operated, will cease to exist. (4)
The scheme has been a hotly debated Indigenous policy issue over
the past 15 years. On the one hand it has been argued that the scheme
provides a cost effective way of providing both community development
and labour market program type objectives for Indigenous Australians,
particularly those without the skills or desire to find non-CDEP
employment or those living in areas with very few non-CDEP labour market
opportunities (Altman & Gray 2005; Altman 2007). On the other hand
it has been argued that the scheme has acted as a disincentive for
participants to find non-CDEP employment and had become a destination
rather than a stepping stone to non-CDEP employment (see, for example,
DEWR 2005; Hudson 2009). Other researchers and commentators have argued
that while CDEP may provide a range of community and social benefits, it
can also act as a disincentive to investing in education and finding
paid employment (Pearson 2007; Hunter 2009).
This paper addresses two research questions. The first is what was
the nature of CDEP employment in 2008 and did the characteristics of
such work change between 1994 and 2008? The second question is how did
the social and economic outcomes of CDEP participants compare to people
in other labour force states (employed outside of CDEP, unemployed and
not-in-the labour force (NILF)) and did these change between 1994 and
2008? The social and economic outcomes examined have been selected to
represent a range of aspects of wellbeing and are those previous
research suggests CDEP may impact on (Altman et al. 2005; Hunter 2009).
These research questions are of interest for several reasons.
First, they are important for interpreting the trends in Indigenous
employment. This is because the Australian Bureau of Statistics (ABS)
has generally classified CDEP participation as equivalent to other forms
of paid employment. Gray and Hunter (2011) find that if CDEP is
classified as employment then there has been little, if any, growth in
the employment rate of Indigenous people, whereas if CDEP is not
classified as employment then there has been strong growth in the
Indigenous employment rate over this period. Whether or not Indigenous
employment increased over this period is critical to assessments made
about the success or otherwise of government expenditure aimed at
increasing Indigenous rates of employment.
Second, the CDEP scheme provides an important example of an
employment program which had direct employment objectives, labour market
program type objectives and community development objectives.
Understanding the economic and social impact of CDEP on participants is
thus of relevance to the design of future labour market programs which
may have both labour market and community development type objectives.
Third, for future comparison, the analysis presented here provides a
useful baseline for CDEP employment and the associated economic and
social outcomes evident just prior to the substantial changes to the
scheme from July 2009.
This paper is structured as follows. The next section provides
information on the changes to the number of CDEP participants since 1994
and describes the characteristics of CDEP jobs and how they have
changed. This is followed by a comparison of social characteristics and
experiences of CDEP participants and those of the non-CDEP employed,
unemployed and those not in the labour force (NILF). The associations
between CDEP participation and a range of social and economic outcomes
is then estimated and compared to the outcomes for the non-CDEP
employed, unemployed and those NILE The paper concludes with a
discussion of the implications for policies that aim to close the gaps
in social and economic outcomes for Indigenous and other Australians.
Number of CDEP participants
The number of CDEP participants and their proportion of the
Indigenous working age population over the period 1997-2010 is shown in
Figure 1. The number of Indigenous CDEP participants has declined from a
peak of around 35,200 participants in 2002-03 to 10,300 participants as
at 30 June 2010. As a proportion of the working age population the
proportion in the CDEP scheme has fallen for men from 1997, while for
women it has fallen from about 2005. (5)
The decline in the proportion of the working age Indigenous
population in the CDEP scheme over the period 1997-2005 is a consequence
of a 'cap' being applied to the number of CDEP places funded.
The decrease in the number of places from 2006 is to a large extent a
consequence of the closing down of urban and regional CDEP schemes.
[FIGURE 1 OMITTED]
Characteristics of CDEP jobs
This section describes the data on the nature of CDEP jobs and
considers the extent to which information is comparable between 1994 and
2008. The analysis is based upon data from the 1994 National Aboriginal
and Tortes Strait Islander Survey (NATSIS) and the 2008 National
Aboriginal and Tortes Strait Islander Social Survey (NATSISS). The 2008
NATSISS and its forerunners--the 1994 NATSIS and the 2002 NATSISS--are
the only large-scale surveys providing national representative samples
of Indigenous Australians which reliably identify CDEP participants. The
NATSIS 1994 and NATSISS 2008 were conducted by the Australian Bureau of
Statistics (ABS) and involved 15,271 and 13,307 respondents
respectively. (6) Unless otherwise indicated, the following analysis is
restricted to the working age population aged between 18 and 64 years.
In the text, the word 'significant' is only used where the
difference between estimates is statistically significant at the five
per cent level.
Given the progressive reduction and ultimate withdrawal of CDEP
from urban and regional areas since 2007, the analysis of changes in the
economic and social outcomes associated with participation in the CDEP
scheme are analysed both for Australia as a whole and specifically for
remote regions.
The characteristics of CDEP jobs examined are: average hours
usually worked each week; under-employment (worker prefers to work more
hours); whether the participant has more than one job (multiple jobs);
the extent to which work allows family and cultural obligations to be
met; and personal income (gross weekly income from all sources). These
characteristics have been chosen because they are important
characteristics of jobs and are available from both the 1994 NATSIS and
the 2008 NATSISS.
Hours worked
The majority of CDEP jobs were, and remain, part-time with average
usual weekly working hours not changing significantly between 1994 and
2008. Average hours worked per week were 23.4 hours in 1994 to 23.8
hours in 2008 (Table 1). According to the program design, CDEP
participants are supposed to work around 15 hours per week in order to
receive the standard CDEP payment. The fact that, on average, CDEP
participants work 23 to 24 hours per week (about 8 hours extra per week)
probably reflects the fact that some participants work extra hours for
CDEP 'top-up' and some participants also held a non-CDEP job.
(7)
The distribution of hours worked by CDEP participants in 1994 and
2008 is shown in Figure 2. The overall distribution of hours worked is
very similar in 1994 and 2008 for both males and females. (8) Between
1994 and 2008, the proportion of CDEP participants working full-time (35
hours per week or more) increased slightly for males (from 14 per cent
to 19 per cent) and was unchanged for females (27 per cent and 28 per
cent in 1994 and 2008 respectively).
Under-employment and multiple jobs
CDEP participants have higher rates of under-employment than other
Australian workers (see ABS 2011). In 2008, about one-third of CDEP
participants indicated that they would like to work more hours (34.4 per
cent). Within the Indigenous population, the rate of under-employment
among CDEP workers was generally lower than that among non-CDEP workers
as whole, although the rate of under-employment for CDEP workers was
similar to that for part-time non-CDEP workers (of whom 39 per cent said
that they would like to work more hours).
Under-employment among CDEP participants, on average, decreased
slightly between 1994 and 2008. The decrease applied to a wide range of
CDEP participants, including males and females, younger and older, and
those in remote and non-remote areas. In contrast, the level of
under-employment among non-CDEP workers was similar in 1994 and 2008. It
is unclear why this difference has occurred.
[FIGURE 2 OMITTED]
There were significant increases in the proportion of CDEP workers
with more than one job for those over 35 years of age, but not for those
under 35 years. The incidence of holding multiple jobs among non-CDEP
workers was only slightly higher than among those in CDEE
This suggests that there may have been some changes in the
composition of CDEP jobs, with the proportion of CDEP participants
holding more than one job having increased and the proportion
under-employed having decreased. The decrease in under-employment is
probably a result of people who wanted to work more hours taking up
increased employment options in the relatively strong labour demand
which occurred over most of the period 1994-2008 (Hunter & Gray
2012a).
Ability to meet cultural obligations
The proportion of CDEP employees who say that working in the CDEP
scheme allows them to meet their cultural obligations is higher than
Indigenous people employed in non-CDEP jobs (in 2008 it was 82.6 per
cent for CDEP participants and 40.8 per cent for those in non-CDEP
jobs). Between 1994 and 2008, the proportion of CDEP participants who
said that they could meet cultural obligations increased from 68.7 per
cent to 82.6 per cent (Table 1). The changes in the proportion of CDEP
participants who indicated that work allowed their cultural obligations
to be met increased in both remote areas and non-remote areas, and hence
cannot be explained by the withdrawal of CDEP from urban and regional
areas.
Personal income
Previous research has found that CDEP participants have somewhat
higher average incomes than people who are unemployed or NILE This is a
consequence of the CDEP 'participants supplement' and the more
generous income test applied to CDEP participants than is applied to
those receiving income support payments. However, CDEP participants have
also been found to have substantially lower income than the non-CDEP
employed.
Information on gross personal income in 1994 and 2008 by labour
force status is provided in Table 2. There are several key points to
take from Table 2. First, as expected, personal incomes of the non-CDEP
employed were much higher than for those in the other labour force
categories. This was true in both 1994 and 2008 and in remote and
non-remote areas. Second, in both 1994 and 2008 the personal income of
CDEP participants was substantially higher than that of the unemployed
or those NILE For example, in 2008 in remote areas the income of CDEP
participants was $359, which was substantially higher than the weekly
income of the unemployed in these areas ($231) and those NILF ($228).
Third, between 1994 and 2008 the weekly income of CDEP participants
in remote areas was basically unchanged at around $360 (in 2008
dollars). (9) Fourth, the income of the non-CDEP employed increased
between 1994 and 2008 in real terms. For example, in remote areas it
increased from $697 to $845 per week over this period (a real increase
of 21 per cent). Given that CDEP incomes have not increased, CDEP income
as a percentage of non-CDEP income has fallen substantially over this
period. For example, in remote areas CDEP income was 51 per cent of the
income of the non-CDEP employed in 1994, and by 2008 it had fallen to 43
per cent.
Duration on CDEP and permanency of CDEP jobs
The 2008 NATSISS collected information on a number of other
characteristics of CDEP jobs which were not collected in the 1994
NATSIS. This section focuses on:
* whether the main part of a participant's job is part of the
CDEP scheme
* the total number of years spent in paid employment (CDEP plus
non-CDEP employment) * the length of time in current job
* whether the participant considers their CDEP job to be a
permanent job, and
* whether the participant has been on CDEP for more than two years.
Historically CDEP participants have been more likely than other
Indigenous workers to be male, relatively young and heavily concentrated
in remote areas (Altman et al. 2005). Consequently, these
characteristics are disaggregated by gender, age and geographic
remoteness (Table 3). On average, CDEP workers have spent more than 10
years in the workforce, a not inconsiderable level of experience. The
only group for which this was not the case was the under 35 year olds
who had an average of 6.5 years in paid employment. Younger non-CDEP
workers also tended to have similarly lower levels of experience. This
is explained by the fact that younger people have obviously had less
opportunity to gain experience. Hence the observation that CDEP workers
had about five years less experience in paid employment than non-CDEP
scheme workers appears largely to be due to the younger age profile of
CDEP workers.
The pattern of time spent in current job was consistent with the
overall experience in paid employment. That is, those CDEP workers with
longer overall workforce experience tended to have had a longer duration
in their current job. However, some groups had had a greater proportion
of their labour market experience in the current (CDEP) job. Younger and
female participants, especially those living in remote areas, seemed to
be more reliant on CDEP scheme work as part of their overall workforce
experience than comparable groups of non-CDEP scheme workers (see Hunter
& Gray 2012b).
The NATSISS 2008 data indicates that two-thirds of CDEP
participants considered their job to be permanent. While it is not
entirely clear how to interpret the term 'permanent' here, in
nearly all categories over 50 per cent of CDEP workers had been in their
current CDEP scheme job for over two years (Table 2). (10) To put this
in context, an even higher proportion of non-CDEP workers considered
their job to permanent (more than 80 per cent).
Occupations of CDEP participants and the non-CDEP employed
A lack of data to date has meant that there has been only a very
limited analysis of the occupations of CDEP participants and how they
compare to the occupational distribution of the non-CDEP employed. The
Census, while providing detailed information on occupation, does not
reliably identify CDEP employment across Australia and so it has not
been possible to obtain reliable national estimates of the occupation
distribution of CDEP participants. The 1994 NATSIS and 2002 NATSISS did
not provide information on occupation. However, for the first time, the
2008 NATSISS provided detailed information on occupation which can be
used to estimate the occupational distribution for CDEP participants and
the non-CDEP employed.
Comparing the occupations of CDEP participants and non-CDEP
employed is one way of understanding the substantial differences in the
nature of CDEP and non-CDEP employment (Table 4). Overall, CDEP
participants were more likely to be working in lower skill level
occupations than the non-CDEP employed
(Table 3). For example, half of male CDEP workers were labourers
(50.4 per cent) compared to a quarter of Indigenous men in non-CDEP
employment who were working as labourers (24.6 per cent). Nonetheless,
there was little difference in the proportion of CDEP participants and
Indigenous non-CDEP employed who are in professional occupations.
One way to summarise this difference in occupational distribution
is to use a segregation index which measures the degree to which the
occupational distributions differ on a scale of zero (no difference) to
100 (maximum difference). Table 4 provides estimates of the segregation
index comparing the occupational distributions of Indigenous males and
females in CDEP to those in non-CDEP employment and to those of the
non-Indigenous employed.
The segregation indexes clearly illustrate the extent of the
difference in the occupational distribution of the CDEP participants and
the non-CDEP employed. The segregation indexes comparing Indigenous to
the non-Indigenous are substantially higher for CDEP participants than
for the non-CDEP employed. For example, the segregation index indicates
that 41.9 per cent of Indigenous males participating in the CDEP scheme
would have to change occupations to be employed in similar sorts of jobs
as non-Indigenous workers. In contrast, 26.2 per cent of Indigenous
males in non-CDEP employment would have to change jobs to have the same
occupational distribution as non-Indigenous workers. The segregation
indexes are lower for Indigenous women than Indigenous men, indicating
that fewer women would have to change occupation to achieve the same
occupation distribution as non-Indigenous women than men would to change
occupation to achieve the same occupational distribution as
non-indigenous men. For example, 33.8 per cent of Indigenous women
participating in the CDEP scheme would have to change occupation in
order to have the same occupational distribution as non-Indigenous
women, compared to 41.9 per cent of Indigenous men having to change
occupation in order to have the same distribution as non-Indigenous men.
Selected social characteristics of CDEP participants and community
development
This section focuses on a range of social characteristics of CDEP
participants and how the CDEP employed compare to the non-CDEP employed,
unemployed and those NILF (as shown in Table 5). Because by 2008 the
vast majority of CDEP participation was in remote areas, the analysis in
this section is restricted to people living in remote areas.
Two sets of social characteristics are examined. The first set
relates to participation in cultural activities and connection to
culture. The specific measures analysed are whether the individual
speaks an Indigenous language, lives on homelands/traditional country
and/or engages in hunting and gathering. These variables are interesting
because, historically, a feature of CDEP has been that the scheme gives
the flexibility and time for participants to engage in customary
cultural activities. In some cases, CDEP activities can involve
customary activities including land management practices. Other
variables are whether respondents are studying and the extent of their
input into community decisions on important issues.
The second set of variables relates to the experience of
discrimination: whether respondents felt discriminated against in the
previous 12 months and whether respondents felt discriminated against in
the previous 12 months because of work (or in the process of looking for
work). Given that the vast majority of CDEP participants are Indigenous,
it is possible that CDEP participants experience lower levels of
discrimination than Indigenous workers in non-CDEP jobs (because there
is a relatively low likelihood of inter-racial interactions in CDEP
workplaces). Several studies have found that some CDEP participants find
CDEP employment is within their 'comfort zone' (for example,
Smith 1994; ATSIC 1997). Finally, the extent to which the participants
had little or no say within their community on important issues is
analysed.
A higher proportion of female CDEP participants spoke an Indigenous
language (72 per cent) compared to Indigenous women in the other
categories in remote areas (Table 5). The unemployed and NILF were less
likely to speak an Indigenous language than the CDEP participants (50
per cent and 53 per cent respectively), but more likely than the
non-CDEP employed (39 per cent). The pattern of results is similar for
Indigenous males.
There is a similar pattern of results for living on traditional
country (or homelands) as for speaking an Indigenous language. Over 60
per cent of Indigenous women participants in CDEP lived on traditional
country compared to around 40 per cent of other Indigenous women living
in remote areas. For Indigenous men, CDEP participants were less likely
to live on country than Indigenous women participants in CDEP (54 per
cent) but they were still more likely than other Indigenous men to live
on traditional country.
Similarly CDEP participants were highly likely to engage in hunting
and gathering, with over four-fifths of women and men being involved in
this form of non-market household production (82 per cent and 86 per
cent respectively). While hunting and gathering is the norm in remote
areas, women and men in other labour force categories tended to be
significantly less likely to engage in this activity. The non-CDEP
employed were the group least likely to hunt and gather (56 per cent and
76 per cent for women and men respectively). However, it should be noted
that, as a group, engagement in these activities by non-CDEP employed
respondents was not significantly different from that evident for women
and men who are NILF.
A large and substantial number of Indigenous people felt
discriminated against irrespective of labour force status, with over
one-quarter of working age adults reporting discrimination in the
previous 12 months. For both males and females the unemployed were more
likely to report being discriminated against in the last 12-months than
other Indigenous people because they were exposed to more potential
discriminators in the process of job search (Biddle et al. 2013). While
female CDEP participants were less likely to experience any
discrimination than other Indigenous females, male CDEP participants
were actually significantly more likely than NILF males to be
discriminated against in the previous 12 months (29 per cent and 20 per
cent respectively).
Another useful comparison is between the discrimination reported by
CDEP and non-CDEP workers. For females, non-CDEP employed are more
likely to report any discrimination than CDEP participants (33 per cent
and 22 per cent); a substantial proportion of this differential is due
to the different rates of reporting discrimination because of work (8
per cent and 4 per cent). In contrast there is no significant difference
in the male rates of reporting discrimination between CDEP and non-CDEP
employed, with around 30 per cent reporting any discrimination in the
previous 12 months and less than 10 per cent specifically identifying
labour market discrimination.
Participants in the CDEP scheme were less likely than the
unemployed or those NILF to say that they had little or no say within
their community on important issues. The CDEP and non-CDEP employed were
similarly likely to say that they had little or no say within their
community on important issues.
In terms of the social characteristics of CDEP participants, the
observations above generally confirm the distinct nature of much CDEP
scheme work, which often involves customary practices and which has been
found in other studies (such as Altman, Biddle & Buchanan 2012).
However, the data only provide weak evidence that CDEP employment was
more consistent with Indigenous preferences than non-CDEP employment
(that is, within some sort of 'comfort zone'). While
discrimination was less evident among female CDEP participants than
among female non-CDEP employed, the reporting of discrimination was not
that different between the respective categories of employed males.
Similarly the sense of efficacy within the community was not
significantly different between CDEP and non-CDEP employed.
The associations between CDEP participation and social and economic
outcomes
The association between CDEP employment and a range of economic and
social outcomes has been the subject of a number of studies (for
example, ATSIC 1997; Altman et al. 2000; Altman et al. 2005; Hunter
2009). After controlling for basic socio-demographic characteristics,
these studies generally found that CDEP participants have slightly
higher incomes and fare somewhat better on a range of social indicators
than the unemployed or NILF Indigenous people, but that participants
have considerably poorer outcomes compared to the non-CDEP employed.
These studies have mostly been based on the 1994 NATSIS, 2002 NATSISS
and census data. This paper is the first analysis of this question using
the 2008 NATSISS.
Empirical approach
It is very difficult (or impossible) to identify the causal impacts
of participation in the CDEP scheme on the wellbeing of participants.
This is because we do not know what their wellbeing would have been if
they had not been participating in the scheme (that is, the
counterfactual). The approach taken in this paper (and other studies,
such as Hunter 2009) is to compare the wellbeing of CDEP participants
with people in other labour force states.
The associations between labour force status (CDEP, non-CDEP
employed, unemployed and NILF) are estimated using multivariate
regression models. The regression models allow the associations between
labour force status and wellbeing to be estimated while holding constant
the effects of other variables which might impact upon wellbeing
independent of labour force status. Several categories of outcomes are
analysed. The first set relates to crime and safety: whether respondents
have been arrested, whether they live in a violent neighbourhood and/or
whether they have been a victim/survivor of crime. The second set of
variables comprises financial measures: whether respondents have
experienced financial stress and/or have a low household income. The
third set relates to health: whether respondents have a disability
and/or whether they have fair or poor health. The final outcome modelled
is whether a person was currently studying. These outcome measures of
wellbeing are all coded as binary variables (zero/one) and thus the
logistic statistical model is appropriate.
This analysis extends and updates estimates of the associations
between labour force status and wellbeing using the 2002 NATSISS (Hunter
2009). The only differences in the measures of wellbeing analysed here
to those analysed by Hunter (2009) are that the earlier study examined
substance use, about which data are not available from the 2008 NATSISS
data; and this paper includes a measure of living in a low income
household a measure not examined in the 2009 study.
The explanatory variables included in the regression model have
been selected to match as closely as possible to those used by Hunter
(2009), in order to compare the extent to which the associations between
CDEP participation and the various measures of wellbeing changed between
2002 and 2008.
The relationship between labour force status and the measures of
wellbeing are captured by including labour force status as an
explanatory variable (specified using a set of dummy variables for being
CDEP employed, non-CDEP employed, unemployed and NILF). A range of other
explanatory variables are included in the model to control for human
capital, demographic, early life experience and other factors which may
impact on wellbeing independent of labour force status. Other
explanatory variables included in the regression models are age (and age
squared in order to capture any non-linear relationship between age and
wellbeing), sex, highest level of educational attainment, whether
respondent lives in a household with both Indigenous and non-Indigenous
residents (mixed household), whether he or she has a partner, whether
there are children resident in household (and the interaction between
having a partner and children being resident), whether the household has
more than one family resident in dwelling, whether the respondent was
taken from family as a child (stolen generation), proficiency in
English, whether the respondent speaks an Indigenous language, their
highest level of educational attainment and geographic remoteness by
State of residence. Details of the explanatory variables are provided in
Appendix Table A.1.
Regression results
The non-linear nature of the logistic regression models means that
interpreting the magnitude of the coefficient estimates is difficult. In
this paper we illustrate the estimated impact of being in each labour
force state on the respective outcome measures, using marginal effects.
The marginal effects show the change in the predicted probability of
social or economic outcomes of moving from unemployment to the other
labour force states (CDEP employed, non-CDEP employed or NILF). The
predicted probabilities are calculated holding constant the value of the
other explanatory variables at the sample means. Table 6 shows the
estimated marginal effects and the base probability, which is the
predicted probability of each outcome variable for an unemployed person
with the average characteristics in the sample. Summary statistics for
the regression models are provided in Appendix A. The full regression
results are reported in Hunter and Gray (2012b).
There are statistically significant differences in the social and
economic outcomes of CDEP participants compared to the unemployed and
those NILE The CDEP employed are 5.8 percentage points less likely to
have been arrested than the unemployed, 5.5 percentage points less
likely to have been a victim of actual or threatened physical violence
than the unemployed, 6.3 percentage points less likely to have
experienced financial stress than the unemployed, 26.6 percentage points
less likely live in a low income household than the unemployed and 6.8
percentage points less likely to report having fair or poor health
status than the unemployed.
These findings--that CDEP participants have higher levels of
wellbeing than the unemployed--are largely consistent with the estimates
of Hunter (2009) using the 2002 NATSISS. The main difference is that in
examining the 2008 data, there are no statistically significant
differences between CDEP participants and the unemployed in terms of
studying, living in a violent neighbourhood, and having a disability. By
contrast, Hunter (2009) found that in 2002 the differences between CDEP
employed and the unemployed were statistically significant for those
outcomes. It is possible that the changes between 2002 and 2008 reflect
a change in the nature of CDEP work. If CDEP was becoming more like a
mainstream work-for-the-dole scheme or a standard employment program and
less like a community development program, then one might have expected
that outcomes would align more with those for the unemployed and NILF
categories. It may also be a reflection of changes in the types of
people in CDEP employment between 2002 and 2008. (11)
While CDEP participants have better outcomes on a range of economic
and social measures than do the unemployed, CDEP participants have much
worse outcomes than the non-CDEP employed for all of the social and
economic outcomes analysed. The differences between the CDEP employed
and the non-CDEP employed are much larger than between the CDEP employed
and the unemployed. For example, the CDEP employed are 5.3 percentage
points less likely to have been arrested than the unemployed, but the
non-CDEP employed are 13.1 percentage points less likely to have been
arrested than the unemployed.
Conclusion
After a quarter of a century of expansion of the CDEP scheme, the
number of CDEP participants has been reduced since 2002-2003 and the
scheme has been gradually withdrawn from urban areas and regional
centres. Indeed, the scheme operated only in remote areas after
mid-2009, and from June 2013 the CDEP scheme will be incorporated into
the new Remote Jobs and Communities Program. The Remote Jobs and
Communities Program will bring together the CDEP, Job Services
Australia, Disability Employment Services and the Indigenous Employment
Program into a single integrated service. This will mean that CDEP will
cease to exist, at least as it has historically operated.
This paper has described the nature of CDEP employment in 2008 and
the extent to which it changed between 1994 and 2008. The paper also
compares a selection of economic outcomes of CDEP participants with
those of Indigenous people in non-CDEP employment, who were unemployed
or who were NILE The characteristics of CDEP jobs in 2008 were very
similar to those found in 1994 and the overall conclusion is that there
has been little change in the fundamental nature of CDEP jobs since
1994. It has remained largely part-time employment, with virtually no
change in the hours worked by CDEP participants. Correspondingly, the
incomes of CDEP participants (in real terms) hardly changed over this
period. While the incomes of CDEP participants were higher than those of
the unemployed and those of people who were NILF, the relativity of the
income of CDEP participants to that of the unemployed and those NILF was
similar in 2008 and 1994. However, over this period the income of the
non-CDEP employed increased more rapidly and hence the incomes of CDEP
participants fell relative to the incomes of the non-CDEP employed.
The general lack of change in the characteristics of CDEP jobs or
the outcomes associated with being a CDEP participant since 1994 is an
important finding in the context of significant changes to both the
number of participants and geographic coverage of the scheme over this
period, and some changes to the rules of the scheme. One possible
interpretation of this is that the CDEP scheme has been very resilient
to attempts by policy makers to alter the way in which the scheme
operates.
Overall, the CDEP scheme appears to give some support for
Indigenous language and customary practice by providing economic
activity that allows participants to live on or near their traditional
country. However, the evidence that participation in CDEP improves
community development through reducing discrimination or enhancing a
sense of personal efficacy in important community issues is weak.
Indeed, having a job is the most important thing for enhancing the sense
of efficacy in the community rather than whether or not one's job
is associated with the CDEP scheme.
While the data presented in this paper are consistent with the
hypothesis that CDEP participation has some small positive social and
health impacts, it equally could be the case that the slightly better
social outcomes for CDEP participants compared to the unemployed are
because those who participate in CDEP have better outcomes prior to
commencing on CDEP (that is, there may be selection effects on
unobservable characteristics of individuals). While it is not possible
to disentangle these alternative hypotheses using the available data, it
is the case that the CDEP employed have only slightly better outcomes
for most measures than the unemployed and generally much worse than for
the non-CDEP employed.
These findings confirm the findings of earlier research that CDEP
positions are very different both in terms of 'job
characteristics' and the social and economic outcomes associated
with participation, compared to non-CDEP employment and therefore should
not be treated as equivalent to mainstream employment.
Whether or not CDEP is treated as employment is not just of
academic interest, but also crucial to assessing whether Indigenous
employment has increased since the mid-1990s. As Gray and Hunter (2011)
demonstrate, treating CDEP as employment leads to the conclusion that
since the mid-1990s there has been little increase in Indigenous
employment, whereas when CDEP is excluded from the definition of
employment then the conclusion is that there has been a substantial
increase in Indigenous employment.
The effect of recent policy changes depends on the extent to which
former CDEP participants secure mainstream employment or move into other
labour force categories. The worst case scenario is that all former CDEP
participants substitute directly into unemployment, other income support
arrangements or even total dependency on other people. In that case,
individuals and associated households will probably experience a drop in
income and encounter smaller impacts on health and social outcomes.
However, some CDEP participants will find mainstream employment and for
these participants there is likely to be a substantial increase in
income. The net effect of the policy change is an empirical question
that is yet to be answered. In the end the net impact of the demise of
the CDEP scheme on social and economic outcomes will depend upon the
effectiveness of the new Remote Jobs and Communities Program, other
government policies and programs and the extent to which Indigenous
communities are able to create and take advantage of new labour market
opportunities.
While the CDEP scheme will cease to exist from mid-2013 in its
current form, it appears that aspects of the scheme will live on in the
new Remote Jobs and Communities Program. The analysis in this paper
provides a benchmark against which the outcomes associated with the
Remote Jobs and Communities Program can be compared.
Appendix Table A1: Summary statistics for logistic regression (%),
NATSISS 2008
Dependent variables
Variable Description Mean Standard
deviation
Arrested Arrested in the previous 5 16.5 37.1
years
Violent neighbourhood Aware of community 35.3 47.8
problems with violent
behaviour (e.g. family
violence, assault & sexual
assault)
Victim Victim of physical 26.1 43.9
violence in the previous
12 months
Financial Stress Whether could raise $2000 47.9 50.0
cash within 2 weeks
Low household income Household income was less 36.8 48.2
than half of the median
Australian income (i.e.
poor')
Disability Has a severe or profound 8.1 27.4
disability
Fair or poor health Self-assessed health 23.8 42.6
status is fair or poor
Studying Currently studying 11.9 32.3
Independent variables
CDEP Employed in CDEP scheme 7.3 25.9
non-CDEP employed Employed outside CDEP 48.4 50.0
scheme
NILF Not in the labour force 35.3 47.8
Age Age (in years) 37 13
Age2 Age squared 1533 996
Mixed household Households with both 35.6 47.9
Indigenous and
non-Indigenous residents
Partner Has a marital partner 57.7 49.4
(married or de facto)
Kids Children resident in 54.0 49.8
household
Partkids Interaction of Partners 34.7 47.6
and Child variables
Multi-family household Households have more than 13.1 33.8
one family resident in
dwelling
Taken Individual taken from 9.2 28.9
family as a child
Difficulty in speaking Difficulty in speaking 3.8 19.1
English English
Degree Holds a degree level 5.8 23.4
qualification
Other qualification Other post-school 24.4 42.9
qualification
Year 12 Completed Year 12 11.9 32.4
Year 10 or 11 Highest level of schooling 32.6 46.9
is Year 10 or 11
Speaks an Indigenous Speaks an Indigenous 23.4 42.4
Language language
Male Respondent is male 42.9 49.5
Number of observations 6,449
Notes: Indigenous persons aged 18-64 years. Geography was controlled
for using 12 dummy variables that identified the disaggregated
information of where respondents lived by State and remoteness
category (classified using the standard ARIA categories), aggregated
where necessary to preserve confidentiality by the ABS. Distribution
not reported here to save space but the summary statistics are
consistent with official ABS estimates of geographic distributions.
Source: 2008 NATSISS data (accessed via ABS RADL).
Acknowledgements
This paper was written as part of an Australian Research Council
Discovery Project grant (DP110101879) entitled 'From welfare to
work, or work to welfare: Will reform of an Aboriginal development
program help close the employment gap?'. We are indebted to Julie
Finlayson, Kirrily Jordan, Matthew James, Deidre McNally, Will Sanders,
John Taylor, participants at a Centre for Aboriginal Economic Policy
Research (CAEPR) seminar, the editors and three anonymous referees for
comments on a previous version of this paper. Patrick Carvalho provided
research assistance.
References
ABS (Australian Bureau of Statistics) (2010a) Labour Force
Characteristics of Aboriginal and Torres Strait Islander Australians,
Estimates from the Labour Force Survey, 2010, Cat. No. 6287.0.
--(2010b) National Aboriginal and Torres Strait Islander Social
Survey, Confidentialised Unit Record File Technical Manual, Australia,
1994 Reweighted (Reissue), Cat. No. 4188.0.55.003.
--(2011) National Aboriginal and Torres Strait Islander Survey,
2008, Cat. No. 4720.0.
Altman, J. (2007) 'Neo-paternalism and the destruction of
CDEP', Arena Magazine, 90, August-September, 33-5.
Altman, J., Biddle, N. & Buchanan, G. (2012) 'The
Indigenous hybrid economy: Can the NATSISS adequately recognise
difference?', in B. Hunter & N. Biddle (eds), Survey Analysis
for Indigenous Policy in Australia: Social Science Perspectives, CAEPR
Research Monograph No. 32, Canberra, ANU EPress.
Altman, J. & Gray, M. (2005) 'The economic and social
impacts of the CDEP scheme in remote Australia', Australian Journal
of Social Issues, 40 (3), 399-410.
Altman, J., Gray, M. & Sanders, W.G. (2000) 'Indigenous
Australians working for welfare: What difference does it make?',
Australian Economic Review, 33 (4), 355-62.
Altman, J., Gray, M.& Levitus, R. (2005) 'Policy issues
for the Community Development Employment Projects scheme in rural and
remote Australia', CAEPR Discussion Paper No. 271, Canberra, CAEPR,
ANU.
ANAO (Australian National Audit Office) (2009) Evaluation of the
Community Development Employment Projects (CDEP) Program Evaluation
Report, Canberra, Department of Finance and Deregulation.
ATSIC (Aboriginal and Torres Strait Islander Commission) (1997)
Evaluation of the Community Development and Employment Program: Final
Report, Canberra, Office of Evaluation and Audit.
Australian Government (2012) Remote Jobs and Communities Program:
Jobs and Stronger Communities for People in Remote Australia, Canberra,
Australian Government http://www.deewr.gov.au/Employment/Consultation/
Documents/RJCP_booklet.pdf
Biddle, N., Howlett, M., Hunter, B. & Paradies, Y. (2013)
'Labour market and other discrimination facing Indigenous
Australians', Australian Journal of Labour Economics, 16 (1):
91-113.
Department of Employment and Workplace Relations (DEWR) (2005)
Building on Success: CDEP Discussion Paper, Canberra, DEWR. Duncan, O.
& Duncan, B. (1955) 'A methodological analysis of segregation
indexes', American Sociological Review, 20 (2), 210-17.
Gray, M. & Hunter, B. (2011) 'Recent trends in Indigenous
labour force status: Re-establishing employment as a social norm?',
Topical Issue 2011/9, Centre for Aboriginal Economic Policy Research,
Canberra, The Australian National University.
Gray, M., Hunter, B. & Lohoar, S. (2012) 'Increasing
Indigenous employment rates', Closing the Gap Issues Paper No. 3,
Canberra, Australian Institute of Health and Welfare and the Australian
Institute of Family Studies.
Hudson, S. (2008) CDEP: Help or Hindrance? The Community
Development Employment Program and its Impact on Indigenous Australians,
St Leonards NSW, Centre for Independent Studies.
Hunter, B. (2009) 'A half-hearted defence of CDEP
scheme', Family Matters, 81, 43-54.
Hunter, B. & Gray M. (2012a) 'Determinants of Indigenous
labour supply following a period of strong economic growth',
Working Paper 81, Centre for Aboriginal Economic Policy Research,
Canberra, The Australian National University.
--(2012b) 'Continuity and Change in the CDEP Scheme',
Working Paper 84, Centre for Aboriginal Economic Policy Research,
Canberra, The Australian National University.
Miller M. (Chairman) (1985) Report on the Committee of Review of
Aboriginal Employment and Training Programs, Canberra, Australian
Government Publishing Service.
Pearson, N. (2007) 'Stuck on the welfare pedestal', The
Australian, 10 February.
Smith, D.E. (1994) '"Working for CDEP": A case study
of the Community Development Employment Projects scheme in Port Lincoln,
South Australia', Discussion Paper No. 75, Centre for Aboriginal
Economic Policy Research, Canberra, The Australian National University.
Spicer, I. (1997) Independent Review of the Community Development
Employment Projects (CDEP) Scheme, Canberra, Office of Public Affairs,
ATSIC.
Endnotes
(1.) An overview of the CDEP scheme and its history up until 2005
is provided by Altman, Gray and Levitus (2005). The changes to the CDEP
scheme in the late 1980s followed a review completed in 1985 (the Miller
Report 1985) and the changes in the late 1990s followed the Spicer
Review (Spicer 1997).
(2.) However, from July 2009 the rules have changed so that while
new CDEP participants may still engage in part-time work in addition to
their CDEP commitments, their CDEP payment is reduced as earned income
increases at a higher rate than was the case previously and it is now
consistent with the standard income test for income support payments.
(3.) According to DEWR there were around 37,000 CDEP participants
in 2005 (DEWR 2005). This number includes non-Indigenous CDEP
participants. It appears that there were around 1,200 non-Indigenous
participants in 2004 at the time when responsibility for CDEP was
transferred from Aboriginal and Torres Strait Islander Social Services
(Aboriginal and Tortes Strait Islander Commission data for June 2004).
The numbers of CDEP participants reported in this paper are an estimate
of the number of Indigenous participants. The estimate of the number of
CDEP participants from the 1994 NATSIS and 2008 NATSISS differ from the
administrative data because of the differences in method of producing
the estimate (for example, social survey and administrative data).
(4.) Grandfathered CDEP participants will continue to receive CDEP
payments until 2017 provided that they remain on the Remote Jobs and
Communities Program.
(5.) Consistent with the decrease in number of CDEP participants
since 2006 (see Figure 1) the proportion of the working age Indigenous
population who are in the CDEP scheme declined between 1994 and 2008,
although the fall is mostly contained to men (11.4 per cent compared
with 7.1 per cent) with little change in the proportion of Indigenous
women working in the CDEP scheme (4.9 per cent compared to 3.9 per
cent).
(6.) Overviews of these surveys are provided in ABS (2010b, 2011).
(7.) The NATSISS 2008 and NATSIS 1994 collected data on hours
worked in all jobs.
(8.) The incidence of holding multiple jobs among non-CDEP workers
is only slightly higher than people employed in the CDEP scheme.
(9.) The lack of real increases in the income of CDEP participants
is a consequence of the linking of government funding for CDEP
organisations to income support payment rates via the 'average per
participant rate'.
(10.) The only exception is non-remote areas where people have a
greater number of labour market alternatives.
(11.) A further difference is that, using the 2002 data, Hunter
(2009) found that CDEP participants were more likely to be living in a
violent neighbourhood than were the unemployed whereas this study using
the 2008 data finds no statistically significant differences between the
CDEP participants and unemployed on this measure.
Table 1: Selected characteristics of CDEP jobs, 1994 and 2008
1994
Weekly Under- 2+ jobs Work allows
hours employed (%) cultural
(%) obligations
to be met (%)
Gender
Female 23.4 40.3 2.1 68.7
Male 25.4 54.1 2.8 64.9
Age
Under 35 years 24.5 52.2 3.1 66.7
Over 35 years 25.2 45.1 1.6 65.0
Remoteness
Remote 24.8 45.9 2.3 75.7
Non-remote 24.4 62.7 3.5 34.1
All CDEP workers 23.4 40.3 2.1 68.7
2008
Weekly Under- 2+ jobs Work allows
hours employed (%) cultural
(%) obligations
to be met (%)
Gender
Female 22.7 25.0 10.4 83.5
Male 24.5 40.4 6.6 82.0
Age
Under 35 years 23.4 40.1 4.5 79.1
Over 35 years 24.3 27.2 12.6 87.0
Remoteness
Remote 24.4 32.6 7.5 82.4
Non-remote 19.2 48.1 12.3 84.1
All CDEP workers 23.8 34.4 8.1 82.6
Note: Table population is Indigenous people aged 18-64 years who were
employed in the CDEP scheme. The number of CDEP workers answering
the 2008 NATSISS questionnaire in remote areas was 450. In non-remote
areas only 32 CDEP participants identified themselves as such in the
NATSISS data-which represented an underlying population of 1,967.
Accordingly great caution should be exercised in interpreting the
non-remote results as they are less reliable than the remote
estimates. Standard errors, which are estimated using a bootstrap
estimator on replicate weights, are available on request.
Source: 1994 NATSIS and 2008 NATSISS data (accessed via ABS Remote
Access Data Laboratory (RADL)).
Table 2: Personal weekly income by labour force status and
remoteness, 1994 and 2008
CDEP Non-CDEP Unemployed NILF
(A$ 2008) employed (A$ 2008) (A$ 2008)
(A$ 2008)
Remote
1994 358.59 697.10 231.29 228.35
(18.41) (22.80) (8.91) (8.14)
2008 363.50 845.30 208.81 263.79
(17.14) (37.32) (16.02) (10.96)
Non-remote
1994 401.10 723.99 241.13 267.37
(36.33) (19.01) (8.92) (7.61)
2008 364.03 851.73 226.64 305.07
(35.96) (19.55) (12.25) ($7.50)
Notes: The population of this table is Indigenous people aged 18-64
years. The income measure is gross (before tax) weekly personal
income. Income from 1994 has been converted to 2008 dollars using
the Consumer Price Index (CPI). Standard errors, which are estimated
using a bootstrap estimator on replicate weights, are reported in
parenthesis.
Source: 1994 NATSIS and 2008 NATSISS data (accessed via ABS RADL).
Table 3: Characteristics of CDEP jobs, 2008
Main job Total time Time in
is part of spent in paid current job
CDEP (%) employment (years)
(years)
Gender
Female 96.3 10.1 3.9
Male 97.8 12.9 3.6
Remoteness
Remote 97.7 11.9 3.9
Non-remote 93.4 11.4 2.0
Age
Under 35 years of age 98.5 6.5 2.6
Over 35 years of age 95.6 18.5 5.1
All CDEP 97.2 11.8 3.7
Considers it On CDEP
a permanent for over
job (96) 2 years (%)
Gender
Female 68.2 56.9
Male 63.8 60.2
Remoteness
Remote 68.7 61.3
Non-remote 41.6 42.0
Age
Under 35 years of age 57.3 52.5
Over 35 years of age 75.9 67.3
All CDEP 65.5 58.9
Notes: The population of this table is Indigenous people aged 18-64
years. Standard errors, which are estimated using a bootstrap
estimator on replicate weights, are available on request. Standard
errors are not reported here to save space, but the inferences in the
text are based on them.
Source: 2008 NATSISS data (accessed via ABS RADL).
Table 4: Distribution of occupation by CDEP status and gender, 2008
CDEP (%) Non-CDEP
employed (%)
Occupation in main job Female Male Female Male
Managers 3.5 4.0 5.5 5.6
Professionals 16.9 6.5 17.4 10.8
Technicians and Trades Workers 6.0 22.8 3.1 20.9
Community and Personal Service 21.1 7.8 27.1 10.9
Workers
Clerical and Administrative Workers 12.9 2.5 22.3 5.4
Sales Workers 5.8 0.1 8.7 3.3
Machinery Operators And Drivers 1-4 6.0 2.1 18.6
Labourers 32.3 50.4 13.9 24.6
Total 100.0 100.0 100.0 100.0
Segregation between Indigenous and 0.338 0.419 0.204 0.262
non-Indigenous occupations
Note: The table population is 18-64 year old employed indigenous
males and females. Segregation indexes are the standard dissimilarity
indexes (Duncan & Duncan 1955).
Occupation classification is based on the Australian and New Zealand
Standard Classification of Occupations (ANZSCO.
Source: Indigenous data from 2008 NATSISS data (accessed via ABS
RADL). Non-Indigenous occupations based on 2006 Census data (see
Hunter & Gray 2012a). Segregation indexes based on author
calculations.
Table 5: Community development and related characteristics by labour
force status and gender, remote areas 2008
Non-CDEP CDEP Unemployed NILF
employed (%) (%) (%)
(%)
Females
Speaks an Indigenous language 38.6 72.2 49.7 52.5
(3.4) (6.8) (6.4) (3.3)
Lives on traditional country 37.6 62.7 43.0 44.8
(homelands)
(3.1) (6.4) (6.9) (3.2)
Non-market production 56.2 81.6 63.9 61.5
(hunting and gathering)
(3.6) (3.8) (6.0) (3.0)
Felt discriminated against in 32.5 22.0 37.6 23.1
the previous 12 months
(3.0) (4.0) (6.0) (2.6)
Felt discriminated against in 8.0 3.6 22.3 1.8
the previous 12 months
because of work (1.4) (1.9) (4.6) (0.6)
Little or no say within 38.6 43.0 51.5 55.0
community on important issues
(3.1) (4.6) (6.5) (2.7)
Males
Speaks an Indigenous language 41.3 72.0 56.1 60.1
(4.4) (4.9) (6.1) (4-1)
Lives on traditional country 35.9 53.8 42.0 50.2
(homelands)
(3.8) (5.3) (4.7) (5.2)
Non-market production 75.5 85.9 82.3 70.0
(hunting and gathering)
(3.4) (2.7) (4.9) (3.9)
Felt discriminated against in 30.4 28.5 35.6 20.2
the previous 12 months
(3.4) (3.8) (5.9) (2.8)
Felt discriminated against in 9.2 8.8 17.4 2.3
the previous 12 months
because of work (1.9) (2.4) (4.7) (1.5)
Little or no say within 38.4 35.7 55.5 52.6
community on important issues
(3.2) (4.0) (6.1) (38)
Notes: The population of this table is Indigenous females and males
aged 18-64 years. Standard errors, reported in brackets, are
estimated using a bootstrap estimator on replicate weights.
Source: 2008 NATSISS data (accessed via ABS RADL).
Table 6: Marginal effect of CDEP and other labour force categories on
selected social and economic outcomes, 2008
Marginal effect (difference from unemployed)
CDEP (%) Non-CDEP NILF (%) Base
employed probability
(%) for
unemployed
(%)
Arrested -5.8 ** 13.1 *** -5.6 *** 21.1
Violent neighbourhood 3.7 -6.1 *** -7.2 *** 40.0
Victim of crime 5.5 * -6.9 ** -2.3 29.0
Financial Stress -6.3 * -25.1 *** 1.0 59.5
Low household income -26.6 *** -56.0 *** -4.8 * 65.5
Disability 0.4 -1.8 ** 7.9 *** 5.1
Fair or poor health -6.8 ** -12.0 *** 8.5 *** 25.0
Studying 1.9 5.1 *** 0.8 7.4
Note: The asterisks indicate statistically significant differences in
the outcome variable for each labour force state compared to the
unemployed. *** indicates a difference at 1 per cent level, ** a
difference at the 5 per cent level and * a difference at the 10 per
cent level.
Source: Calculated using multivariate regression results reported in
full in Hunter and Gray (2012b).
