Educating oneself out of social exclusion

Providing more education and training is considered one means by which to reduce the extent of social exclusion and consequently has been a key focus in recent public policies.

Using the first ten waves of the Household, Income and Labour Dynamics in Australia (HILDA) Survey as well as data from the Survey of Education and Training, the research builds a multi-dimensional measure of social exclusion comprising: material resources (household income and expenditure); employment; education and skills (literacy and numeracy, educational attainment, work experience); health and disability; social interactions; community (neighbourhood quality, civic participation, volunteerism); and personal safety. The authors are then able to show how social exclusion varies across different levels of educational attainment and over time.

The authors also simulate the effect on the measure of multi-dimensional social exclusion of the Council of Australian Governments (COAG) target: halving, between 2009 and 2020, the proportion of 20 to 64-year-olds without at least a certificate III qualification. This mind experiment takes advantage of the correlations between the various dimensions by assuming that the outcomes of the 'new certificate III graduates' are the same as the 'previous certificate III' graduates. In a sense therefore it is a 'best case' simulation and assumes that the quality of the education expansion induced by the COAG target is high.

Key messages

• The level of social exclusion has declined over the decade beginning in 2001, except during the period around 2008—10, presumably as a result of the Global Financial Crisis.
• Education is a powerful marker of social exclusion. Those who are early school leavers or have a certificate II as their highest qualification suffer from social exclusion to a far greater degree than those with other levels of educational attainment. This is true for all dimensions of the index.
• The impact of improved basic educational levels on social inclusion is potentially very significant; for example, if we calibrate our cut-off of the measure of social exclusion so that around 10% of the population is in the socially excluded category and then conduct the COAG target simulation, the percentage of the population who are socially excluded drops to under 7%.

Notwithstanding its statistical complexity, the research clearly shows the power of attacking poor levels of education to reduce social exclusion.

Tom Karmel
Managing Director, NCVER

Social exclusion is inherently multi-dimensional, with many, at times interconnecting, factors with the potential to impede an individual from fully participating in society. In this report we examine the relationship between education and training and social exclusion in Australia using data from the Household, Income and Labour Dynamics in Australia (HILDA) Survey and the Survey of Education and Training (SET) conducted by the Australian Bureau of Statistics (ABS). The measure of multi-dimensional social exclusion used builds on earlier work for Australia and identifies seven dimensions of exclusion: material resources; employment; education and skills; health and disability; social interactions; community; and personal safety. In each of these dimensions a person is proportionally excluded, with the proportion ranging from 0 (not excluded) to 1 (fully excluded). These proportions are then summed to express multi-dimensional exclusion by a single 'sum-score', which therefore ranges between 0 and 7. A person is deemed to be multi-dimensionally socially excluded if their sumscore exceeds a threshold level. Although less intuitive, the principle is identical to, say, classifying a person as 'poor' when their income falls below a certain threshold level.

Having defined how social exclusion is measured, the report then asks:

• What is the extent of social exclusion in Australia?

We find that a key overarching message from this report is that setting the sum-score's threshold level for being deemed multi-dimensionally socially excluded at 1 or 2 indeed only affects the level of social exclusion in Australia. The story, such as what is happening with exclusion over time, or the relative levels of exclusion among different subgroups in the population is, in the main, independent of such threshold levels.

In terms of the trend in exclusion, we find that the incidence of multi-dimensional social exclusion has been declining from 2001 to 2008, but this downward trend was reversed and exclusion rates increased from 2008 to 2009, and again from 2009 to 2010. For those individuals who are early school leavers holding, at most, certificate II, the incidence of exclusion did not decline over time but was flat to slightly rising from 2001 to 2008. However, this group, too, has experienced a sharp increase since 2008. Although not formally tested, the period since 2008 coincides with the Global Financial Crisis (GFC), which is most likely to be at least partly responsible for the increase in our multidimensional social exclusion measure.

• Are particular education and training qualifications more or less likely to be associated with social exclusion?

There are clear links between education and measured social exclusion. However, while the exclusion rates are lowest for those with the highest levels of education ('higher education' ¹ and [advanced] diplomas) and the exclusion rates for people with Year 12 and those with certificates level III and IV even overlap, there is only a real dichotomy between early school leavers with at most certificate II, and the rest. This suggests that the biggest impact on social inclusion through education is expected to come from efforts to increase Year 12 completion rates and/or completion of certificate level III qualifications rather than from efforts to increase the proportion of people with even higher levels of qualifications.

• To what extent do low levels of education and training, in other words, the education and skills dimension, contribute to social exclusion?

An advantage of our measure of social exclusion is that it can be decomposed and fully apportioned to each of the seven dimensions that make up our measure. That is, the contribution by each of the seven dimensions can be expressed as a share, with the shares over all seven dimensions adding to 100%. Overall, approximately 12% of social exclusion can be attributed to the dimension education and skills. For all education subgroups the material resources dimension (that is, income, expenditure, net worth, financial stress) is the largest driver of exclusion, contributing between 30 and 40%, followed by community. The share of 'community' is stable across education levels, contributing approximately 18% to the overall adjusted headcount. There is also a strong positive relationship between age and the contribution of health, with health contributing about 16% for those aged 45 years and over.

• If the target set by the Council of Australian Governments (COAG) in relation to halving the proportion of the population aged 20—64 without qualifications at certificate III level and above by 2020 is met, how will this impact on the level of multi-dimensional social exclusion?

We simulate the effect of the COAG target to halve the proportion of Australians aged 20—64 without qualifications at certificate III level and above between 2009 and 2020 on our measures of multidimensional exclusion. We consider two impacts: the direct effect and the cumulative effect.

The direct effect is the effect of (randomly) changing education levels in half the number of cases where individuals report Year 11 and/or certificate I/II. The impacts measured as absolute changes in the level of multi-dimensional social exclusion are modest.

The cumulative effect considers the impact of superimposing the COAG target on the analysis, taking into account the multiplier effect of increasing people's education with the resultant better health, higher incomes, higher labour force participation rates etc. This effectively assumes that, by improving the individual's education level, all the characteristics associated with this higher education level are also inherited. Not surprisingly, the impact on the multi-dimensional exclusion measure is stronger in the case of a cumulative impact, with an approximate 30% reduction in the headcount measure of social exclusion. It might be considered that this is mainly due to assuming higher incomes in tandem with assuming higher education levels, but this is not the case, as the income poverty headcount ratio barely changes under the simulation of the cumulative effect. It is the other dimensions, and in particular their combined impact, that reduce multi-dimensional social exclusion under the COAG scenario, assuming a cumulative effect of lifting education levels.

¹ Higher education includes postgraduate degrees (master's or doctorate), graduate diplomas, graduate certificates, and bachelor or honours degrees.