Home | Index |O'view |Glossary | Part 1 | Part 2 | Part 3 | Part 4 | Part 5 | References | Appendix on Method 
	 Acknowledgements | Boxes - 10 | Figures - 11 | Tables - 18 | Disclaimer | Copyright


	 8. Beyond capital augmented NNI 
		a	A stock-take
		b	Non-economic aspects of wellbeing 
		c	Gross National Suffering 

	 9. Non-economic environmental impacts 

	10. Health 
		a	Introduction
		b	Mental Health 
		c	Obesity 

	11. Employment-related life satisfaction 
		a	Introduction 
		b	Job satisfaction, underemployment, overwork and work/life balance 

	12. Political and social capital 
		a	Political capital 
		b	Social capital 

8. Beyond capital augmented NNI 

a	A stock-take 

So far, we have discussed a number of adjustments that we could make to 
GDP to construct a more comprehensive measure of wellbeing. 

1	Using NNI, a measure of the income Australians have for consumption, 
	rather than GDP, a measure of production. 

2	Accounting for changes in those aspects of capital not taken into 
	account in NNI, namely: 
		o the net accumulation of intangible capital; 
		o the net depletion/accretion of Australia’s most valuable natural 
		  capital stocks, subsoil assets and land, in creating that income; 
		o possible reductions in future consumption possibilities arising 
		  from climate change. 

3	Adjusting income growth to take into account the distribution of income 
	across the population. 

As all these elements are enumerated in dollars, though assumptions must be 
made of varying persuasiveness, the magnitude of each indicator is expressed 
naturally in dollars, enabling commensurability between the measures and 
enabling us to aggregate them into a broader measure of wellbeing. 

However, a number of other important aspects of wellbeing are less easily translated 
into a dollar value. 

They include the wellbeing (over and above the wellbeing already captured by NNI in
income that can be spent on these items) generated by 

	1	a healthy environment; 

	2	good health; 

	3	employment-related satisfaction; 

	4	the quality of governance; and 
	5	social capital or people’s sense of community and mutual obligation 
		to one another. 

The following chapters deal with these issues and propose the terms on which 
they are added to our index of general wellbeing. 

However, the first of these items is dealt with in this chapter in which we summarise our
practical reasons for not including it in our index of wellbeing. 

b	Non-economic aspects of wellbeing 

It is no surprise that there are non-economic aspects of wellbeing. 

However, as explained, the anchoring of our index in the national accounts does provide 
us with some anchor with which to calibrate the relative importance of different aspects 
of wellbeing. 

This is an imperfect – indeed biased – way to calibrate  these weightings, but the main
alternative seems much worse. 

For, as we have seen, pure composite indices appear to have made negligible progress
in dealing with the incommensurability of the various aspects of wellbeing, leading 
most of them to simply posit that each aspect is equally important. 

But given the difficulty of making any progress at all on such a difficult problem, it is 
not arbitrary to assume that the amount of resources a democratic polity expends in
various domains – say in health or education – by way of its own private and public 
democratic choices offers some clue as to its relative importance to that population 
in providing for its wellbeing. 

Thus if by a combination of private and public decision making Australians spend x 
per cent of national income on education and y per cent of national income on health, 
this provides a starting point for determining the relative importance Australians 
(considered as a group) give these things in their lives. 

We can then go beyond this as an assumption and make adjustments to the pure 
national accounting measures reflecting our own investigations into their relative 
importance and/or our values. 

This is effectively what we have done above with regard to the distribution of national 
income, and we extend that approach to other non-economic factors of wellbeing. 

In this regard, where we think there are good measures of the quality of the output of 
various sectors (or of the way in which general technical and organisational change 
are contributing to wellbeing) we can add these to our index of wellbeing to adjust 
for any under or over weighting that might otherwise characterise our index. 

In consequence this part of the report deliberates on a range of additions to the index. 

Notably we have not added an adjustment for education, because education inputs 
are strongly represented as a substantial component in NNI and then educational 
outputs are counted again in our measures of the most important capital item in the 
index – human capital.

We are unashamed of the implicit double counting in this regard because of the 
fundamental importance of education and human capital in human wellbeing that goes 
well beyond its economic significance. 

Likewise, as proposed below, health receives similar treatment in our index because,
in addition to being a precondition for economic prosperity, its significance for wellbeing 
goes well beyond its economic contribution to our lives. 

Thus, as is the case with education, income that sustains inputs to health is counted 
once as part of NNI with measures of health outcomes being counted again – although 
in the case of health we do this more explicitly as an element of non-economic wellbeing. 

c	Gross National Suffering 

One thought with which we began this exercise was that it might be possible to introduce
into the scheme of measurement some focus on the direct causes of suffering. 

This has obvious appeal because most of us care far more about the avoidable causes of 
suffering in our lives than we do for the next increment of wellbeing in domains in which 
we regard ourselves as doing relatively well (or do in our more sober moments, or in 
hindsight when we have succumbed to some important setback on our lives). 

We saw some merit in this idea for these reasons: 

1	It is commonsensical, capturing human experience; 

2	It is consistent with one emerging ‘stylised fact’ from behavioural 
	economics – namely that people are deviate from complete ‘rationality’ 
	as framed in economic discourse because they care more about losses 
	than winnings, in economic and other aspects of their lives; 

3	To the extent that such measures are used as a guide for policy it may 
	have salutary effects and uncover ‘low hanging fruit’ from the 
	perspective of promoting Australians’ wellbeing; 

4	In doing so it may offer a route of escape from Easterlin’s paradox. 

	If most people describe themselves as relatively happy, a focus on 
	reducing suffering might produce more meaningful variation in 
	wellbeing than continuing exclusive focus on happiness. 

We agree with Denis Healey citing Kolakowski that, in a prosperous country,one of the main 
tasks of policy should be the task described by Healey as “eroding by inches the conditions 
which produce avoidable suffering” (Healey, 1989, p. 472-3). 

To our surprise we were unable to find this idea well represented in the literature. 

However, when we looked directly at the problem we found that many causes of what is 
clearly suffering of a high order – for instance suicide or even homelessness – were 
sufficiently rare in our community that for them to make much difference to our index 
would require weighting that would be highly contentious. 

Box 8: The difference between strategic planning and an index of wellbeing

It is a commonplace of public policy as it is of business management that what 
gets measured gets done. 

Inspired by this notion, a number of Australian states have been world leaders in a 
process of strategic planning by which the State Government engages with the public
in producing a desired set of outcomes. 

Such outcomes might involve any number of detailed commitments to guide policy in 
seeking specific social outcomes. 

Thus for instance South Australia has over 100 targets in its state plan. 

One is the reduction of road deaths. 

This is a self-evidently worthwhile objective and one that is being realised as demonstrated 
in successive reports on the plan. 

Yet annual road deaths amount to about 100 people in South Australia with trend annual 
changes being fewer than 10 deaths per year. 

However concerning each and every death is, clearly including them in an index of the 
wellbeing of South Australians would simply bury the issue in the mass of other 
determinants of South Australians’ wellbeing. 

Thus, while the process of strategic planning can focus on any worthwhile objective, 
a community index of wellbeing must be more summary and more general in its 

One can argue about the worth of having a single, summary instrument of wellbeing, 
but if we are to have one, it must be parsimonious in its responsiveness to causes for 
fear that the signal of each cause gets lost in the noise of them all. 

Nevertheless, there is a range of sources of suffering that are sufficiently widespread in 
our society that they can have an appreciable impact on our index. 

They are also recognised as sources of community concern. 

In each case we had intended to include them in what had begun as a composite index. 

But as we developed our means of weighting, we thought that a unifying theme for most 
if not all of these areas was their capturing of important sources of avoidable suffering in 
our society. 

Another way of looking at many of the components of this Part Four of the report is to 
envisage them as constituting a summary index of Gross National Suffering. 

9. 	Non-economic environmental impacts 

As noted above, natural resource depletion and land degradation are not the only ways 
in which environmental degradation can reduce our wellbeing. 

1	Air and water pollution can impair health. 

2	Economic development can also impair people’s amenity from the 
	national environment including their valuation of biodiversity.(41)
	The national accounts do not measure these wellbeing effects satisfactorily. 

	A large literature (Nordhaus and Tobin, 1973) documents a variety of 
	approaches to better reflect the impacts of environmental degradation. 

	However, the difficulties in expressing such effects in ways that are
	commensurate with other aspects of wellbeing has meant that no consensus
	method has emerged and we doubt one ever will. 

	Several findings emerge from this literature (see Box 9). 

3	Environmental performance indicators can capture current direct 
	environmental impacts on people and/or indicators of changes in the 
	quality and quantity of future environmental stocks. 

4	Indirect impacts of the environment on wellbeing such as the pleasure 
	obtained from a pristine natural environment are extremely difficult to value 
	and not included in any of the studies listed. 

5	Developed countries such as Australia generally perform well on direct 
	environmental health measures, presumably because of relatively strong 
	regulatory controls on pollutants. 

	This suggests, firstly, that the political process in Western democracies
	substantially ‘internalises’ the direct health costs of environmental degradation 
	through regulation of hazardous emissions. 

	Secondly, by the same token, the health impacts of environmental degradation
	are likely to change slowly.

	Accordingly, including them in our wellbeing index is unlikely to affect its 
	movement significantly. 

Loss of biodiversity may also have direct economic costs, if, for instance, it degrades 
agricultural productivity or resilience.  

Box 9: Existing approaches to environmental accounting

The YaleEnvironmental Performance Index  (EPI) ranks 163 countries on 25
erformanceindicators across 10 policy domains covering both environmental public
health and ecosystem vitality. 

In 2010 Australia scored 91.73 out of 100 for environmental health but only 39.58 for 
ecosystem vitality, ranking 51/163 with an overall score of 65.7.(42) 

This is below most European countries but about the same score as Canada
(66.4, 46th), the United States (63.5, 61st) and Brazil (63.4, 62nd), and significantly 
outperforms developing countries in Africa and Asia. 

The EPI replaces the earlier Environmental Sustainability Index, which was a much
more complex index comprising 76 variables tracking human vulnerability, social and 
institutional capacity and global stewardship as well as current environmental performance.

In a number of areas, the EPI uses ‘distance to target’ indicators that monitor a country’s
performance against agreed environmental benchmarks such as air pollution levels (ie. 
meeting the benchmark would give a country a score of 100/100 for that indicator). 

Ecological Footprint Index  tracks resource demand by calculating the amount 
of land required to produce the biological inputs to commercial production such as 
cropland, grazing, forestry, fishing, as well as land in built-up areas and carbon sinks 
that would be required to offset greenhouse gas emissions. 

This can be compared to resource supply by assessing a country’s total land and 
water resources (bio-capacity) to see if the footprint is sustainable. 

Australia’s ecological footprint has averaged about eight hectares per capita since the 
1960s, but bio-capacity has fallen from 30 hectares per person to about 17 in 
2007, as finite areas are required to service an increasing population.(43) 

The Footprint Index does not include direct impacts of environmental degradation 
on wellbeing, for instance via pollution. 

Environmental Accounts put a monetary value on the natural resource assets 
used in commercial production, with values generally based on the resource 
rents charged for use. 

Examples in Australia include the satellite accounts for energy (ABS, 2009), 
water (ABS, 2010b) and land use. 

The ABS’s experimental estimates to account for changes in subsoil, land and forest 
assets between 1993-94 and 2000-01 (ABS, 2003) show that year-to-year changes in 
environmental capital stocks can be positive as well as negative. 





Box(Cont): Existing approaches to environmental accounting

The Genuine Progress Indicator and other forms of Green GDP make a 
series of adjustments to GDP to account for environmental impacts. 

The Australia Institute’s GPI for Australia subtracts costs calculated for noise 
pollution, irrigation water use, urban water pollution, air pollution, land degradation, 
loss of native forests, depletion of non-renewable energy resources, climate change 
and ozone depletion. 

Collectively these environment costs subtracted $60 billion from Australia’s welfare in 
2000 (Australia Institute, p. 20). 

The ABS MAP Environmental Domain considers six facets of environmental 
progress – biodiversity, land, inland waters, oceans and estuaries, atmosphere 
and waste. 

Headline indicators have been agreed only for biodiversity and atmosphere, but a 
number of secondary indicators are included that are similar to the environmental 
domain indicators of other composite indices such as the Canadian Index of Wellbeing. 

The Yale EPI provides the most comprehensive indicator of the non-health-related 
aspects of environmental degradation in its sub-index of eco-system vitality.(44) 

To track this in our own index we would take this measure – released biennially – 
with the EPI’s measures of climate change removed to prevent double counting 
within our index. 

This could be used for now, at least until the ABS MAP project develops headline 
indicators for a larger number of sub-domains in the environmental space. 

However, weighting this index is problematic. 

If it were to be included it would be difficult to justify giving it a large weight. 

Given this it would have negligible impact on the overall index. 

Further, there is no evidence we can find that the state of eco-system vitality has a
direct impact on human wellbeing. 

For this reason we are collecting and recording this sub-domain index but 
currently giving it a zero weight for the time being. 

Should the index be provided in a form that enabled others to reweight it according 
to their own values and preferences – as proposed below – this would give them 
the means to give the issue greater weight. 


To avoid double counting, the climate change domain of the Yale EPI will be excluded.  

Table 16: EPI 2010 Indicators, weighting and latest data

Ecosystem Vitality (50%) 

1. Climate change (25%) 		5. Forest cover (2.083%), 2005 
   · Greenhouse gas emissions from 	
     land use (12.5%), 2005 		6. Biodiversity & Habitat (4.167%) 
   · CO2 emissions from electricity	   · Biome Protection (2.083%), 2009  	
     generation (6.25%), 2007 		   · Marine Protection (1.042%), 2007
   · Industrial greenhouse gas 		   · Critical Habitat (1.042%), 2005 
     emissions intensity (6.25%), 2005 	
					7. Water effects on ecosystem (4.167%) 
2. Agriculture (4.167%) 		   · Water Quality Index (2.083%),2009			
   · Agricultural water intensity 	   · Water Stress Index (1.042%),1995	
     (0.833%), 2002 			   · Water Scarcity Index (1.042%), 2007
   · Agricultural subsidies (1.25%), 
     2008 				8. Air pollution effects on ecosystem (4.167%) 			
   · Pesticide regulation (2.083%), 	   · Sulphur dioxide (2.083%), 2000 
     2007 				   · Nitrogen oxides (0.694%), 2000 
					   · NMVOCs (0.694%), 2000
3. Fisheries (4.167%) 			   · Ecosystem ozone (0.694%), 2000 		
   · Marine Trophic Index (2.083%),
   · Trawling Intensity (2.083%), 2004 

4. Forestry (4.167%) 
   · Growing stock (2.083%), 2005 

Environmental Health (50%)

1. Environmental Burden of Disease 	3. Water effects on humans (12.5%) 		
   (25%), 2004 				   · Access to water (6.25%), 2006 			
					   · Sanitation (6.25%), 2006 
2. Air pollution effects on humans 
   · Indoor Air pollution (6.25%, 2007 
   · Outdoor Air pollution (6.25%), 2006 

Recommendations for the HALE Index

Issue 			Indicator 			Preferred Weight (%) 

Ecosystem vitality 	Track using Yale EPI Index’s 	Zero 
(other than climate 	Ecosystem vitality measure for 
change) 		Australia, without climate change 

10. Health 

	a	Introduction 
	b	Mental Health 
	c	Obesity 

a	Introduction 

Health is a matter of paramount importance to us all – a prerequisite of human 

About nine per cent of national income is spent, by governments and households, on 
Australians’ health, and this is captured in measures of NNI. 

This provides a rough approximation of the relative importance of health to all Australians, 
although of course those with health difficulties would be prepared to spend vastly more 
than this if it enabled them to substantially improve their health. 

As sympathetic as one might be to such a situation, an index of overall wellbeing must 
aggregate, as best we can, the wellbeing and preferences of all Australians. 

Nevertheless, as is the case with our measures of education, having calibrated the
relevant weighting to be given this domain in our wellbeing index, if appropriate metrics 
can be found, it is preferable to measure outputs rather than inputs. 

In this regard, if one seeks a single, summary measure, it is hard to go past life expectancy
at birth as a measure of the overall health of a population. 

-This measure is used for both the UN’s Human Development Index and the OECD’s Better 
Life Index. 

As a developed country, Australia has a high life expectancy that has continued to slowly 
increase over recent years. 

However, life expectancy alone does not tell us about people’s state of health 
while they are alive. 

This may be measured in either subjective or objective ways. 

Subjective measures of self-reported health status are included in the National Health 
Survey conducted every three years (last conducted in 2007-08). 

Objective measures include the burden of disease calculated by the Australian 
Institute of Health and Welfare as well as hospitalisation rates. 

An alternative way of measuring improvements in health is the proportion of deaths 
and serious injuries that are preventable. 

Preventable health events include vaccine-preventable conditions, chronic conditions
that can be managed through lifestyle interventions such heart disease, asthma, diabetes 
and anaemia as well as acute conditions such as dehydration or dental conditions. 

The AIHW includes annual data on the rate of potentiall preventable hospitalisations as 
part of its Australia’s Hospitals publication. 

In 2009-10, 8.1 per cent of all hospital admissions were for preventable conditions. 

This is equivalent to 30.1 preventable admissions per 1,000 (age-standardised) population 
(AIHW, 2011). 

b	Mental health 

In addition, as the AUWI data demonstrates, mental illness has a powerfuleffect on 

The index demonstrates this most particularly of those suffering from such conditions but 
that unhappiness must also radiate out from the direct sufferer to family members (Fadden 
et al., 1987). 

Given its dramatic effect on wellbeing (see Figure 8) and the possibility that policy can 
substantially improve it, we include it in our index of wellbeing. 

Figure 8: The impact of major medical conditions on wellbeing 
Source: AUWI, 2010, p. 30 

National data on the prevalence of mental illness was last collected in the ABS 2007 
National Survey of Mental Health and Wellbeing. 

At that time 3.2 million, or 20 per cent of the adult population, reported experiencing a 
mental health disorder in the last 12 months. 

Of these, only one-third had accessed medical services to assist them manage their disorder. 

Given the impact on wellbeing of mental illness is likely to be mediated by the effectiveness 
of any treatment a person receives, we recommend the index include a measure of untreated 
mental illness rather than all mental illness.(45)

The COAG National Healthcare Agreement includes two measures of progress for addressing 
mental illness. 

These are: 

1	the proportion of the population receiving clinical mental health services; and 

2	the proportion of people with mental illness who have a GP treatment plan. 


 Ideally we would want to adjust the quantity of treatment for its quality or effectiveness. 

However, we have been unable to find sufficiently detailed data to allow us to do this.  

While the former measure provides a more comprehensive measure of health services 
provided for people with mental illness (both acute interventions and ongoing support), 
we expect the latter provides a closer approximation of how well mental illness is
managed, reducing its negative impact on wellbeing and accordingly use this as our 
measure of the alleviation of avoidable suffering owing to poor mental health.(46) 

c	Obesity 

The AUWI data also suggests that moderate and severe obesity (47) takes a substantial
toll on wellbeing. 

Mildly obese people are also less happy than normal and overweight people, although 
the reduction in life satisfaction is not outside the so-called ‘normal range’ of happiness. 

Given the prevalence of obesity – affecting almost 25 per cent of the population in 2007-08, 
up from about 20 per cent in 2001– its clear impact on wellbeing (see Figure 9) and the 
relative ease with which it can be measured using ABS data, we include obesity rates 
as a factor in the index. 

Data on the percentage of people with mental illness with a GP treatment plan is available 
only for 2007-08 and 2008-09. 

Values for earlier years are extrapolated based on the rate of GP mental health 
consultations per 1,000 population, from AIHW (2010) Mental Health.  

The NHMRC and WHO organisation guidelines suggest an adult is obese if their Body Mass 
Index (weight in kg divided by the square of height in metres) is 30.0 or greater. 

The ABS in its National Health Survey also adopts this definition. 

The AUWI further splits obesity into mild obesity with a BMI of 30.0 to 34.9, moderate 
obesity as a BMI of 35.0 to 39.9, severe obesity as a BMI of 40.0 to 44.9 and very severe 
obesity as a BMI of 45 and over. 

Data from the ABS does not allow us perform such a split.  

Figure 9: Bodyweight and happiness
Source: AUWI, 2010, p. 31

Accordingly for the health domain we recommend constructing an index that takes into 
account not only overall health and longevity, but also how well Australia is doing in 
preventing avoidable health problems.

This index is weighted to be equal to the current spending on health as a percentage 
of GDP (currently 9 per cent). 

As well as this we deduct from our index of wellbeing amounts estimated to approximate 
the negative effect on wellbeing of mental illness and obesity. 

Health domain of the HALE Index
		Indicator 				Narrative 				

Physical	1   Life expectancy at		Life expectancy increases very slowly over time
		2   rate of potentially		The rate of preventable hospitalisations have 
		    preventable 		decreased over the period 

Mental 		Treatment rates, 		While the treatment rates for mental 
		proxied by the 			illness have improved, this has been 
		percentage of people 		outstripped by growth in the number of
		with mental illness 		people with mental illness 
		with GP treatment 
		plans, from National 

Obesity 	Proportion of adult 		Obesity rates have increased over the
		population measured 		period
		as obese, from the 
		ABS National Health 
11.	Employment-related life satisfaction 
		a	Introduction
		b	Job satisfaction, underemployment, overwork and work/life balance

a	Introduction

The impact of unemployment and underemployment on reduced economic activity and 
consumption is already captured in NNI. 

Similarly, the atrophy of human capital stocks due to long-term unemployment is included 
in our calculations of changes in economic capital. 

However, the literature suggests that there are other non-economic links between 
unemployment and wellbeing. 

Unemployed people are significantly more likely to suffer poor psychological health such 
as anxiety, depression and behavioural problems (Cole et al. 2009).(48)

In Australia, unemployed respondents to the Australian Unity Wellbeing surveys typically 
rate all dimensions of their life satisfaction about seven to ten percentage points lower 
than the general population. 

Note that this very large disparity is also reflective of a range of other factors. 

Thus unemployed people tend to come from lower in the income scale (even while they 
are employed) and thus come from a population with a lower self-reported wellbeing than

Further, even if we could assume that someone’s lack of employment, mental or 
physical illness or disability caused reduced wellbeing (rather than causation running the 
other way) these conditions are disproportionately shared by those with low wellbeing. 

So observing unemployed people at seven to ten per cent lower wellbeing does not 
account for the extent to which that lower wellbeing might be driven by co-morbidities. 

That having been said, other studies both in Australia (Headey and Wooden, 2004 and 
Carroll, 2007) and overseas (49) confirm a strong relationship between unemployment 
and unhappiness, even after accounting for other factors that may also affect wellbeing 
such as health and marital status, through a ‘fixed effects’ model specification. 

Two recent ‘fixed effect’ studies of the impact of unemployment on wellbeing using 
Australia’s HILDA data set suggest being unemployed reduces life satisfaction by about 
1.6 percentage points in the year it first occurs (Wooden et al. 2009), falling to 0.8 
percentage points as people adapt over time (Fritjers et al., 2010). 

This suggests that only about 20 per cent of the gross wellbeing loss among the 
unemployed observed in the AUWI survey is attributable to unemployment alone, 
rather than other factors such as low income or poor health. 

 Note, there is some prospect for double counting here between accounting for
employment related life satisfaction and mental health. 

However, our measures of mental health focus on treatment, whereas the measure here 
focuses on the creation of conditions that are conducive to poor mental health. 

This reduces, though it does not eliminate, the amount of double counting.  

See Winkelman and Winkelman (1998) and Gordo (2006) for German data, Clark (2003) 
for UK data and Blanchflower and Oswald (2004) for US data.  

To allow for these co-morbidities, we impose limits on the extent to which any one 
condition is taken to reduce wellbeing by taking the fixed effects values from the 
HILDA studies. 

Even having done so, it appears that the impact on wellbeing is likely to considerably 
outweigh the income that would have been earned by the person should they have been 

For example, the value of lost wellbeing for the 5.22 per cent of Australians 
unemployed in June 2010 would be worth at least $7.15 billion, or 0.7 per cent of NNI,
based on Fritjers calculation that an unemployed person would need to receive a one-off 
payment of $11,500 to compensate for the negative wellbeing impact of unemployment 
that they do not adapt to over time. 

Negative health and other outcomes appear to worsen the longer a person remains 
unemployed (Cole et al., 2009).(50) 

Further, though there are ‘adaptation’ effects that tend to reduce the negative wellbeing 
impact of unemployment on the unemployed, others argue that people do not adapt to 
being unemployed over time (Winkelman and Winkelman, 1998). 

Further, there is some evidence of a ‘scarring’ effect such that people, once they have 
been unemployed for some substantial period of time, never return to the higher wellbeing 
they had before becoming unemployed, even when they go back to work (Lucas et 
al., 2004; Cole et al., 2009). 

For example, Clark et al. (2001) found wellbeing is lower not only for the current 
unemployed, but also for those with higher levels of past unemployment. 

Men who have been unemployed for roughly 60 per cent of their time in the labour force 
over the past three years are indifferent (in terms of life satisfaction) between current 
employment and unemployment. 

This suggests a scarring effect. 

b	Job satisfaction, underemployment, overwork and work/life balance 

There is less literature on the impact of job satisfaction, underemployment or 
overwork on life satisfaction. 

It appears that employees with low levels of job satisfaction (51) or who feel over or 
underworked, report lower rates of subjective wellbeing than people who enjoy their 

Some studies have also found that moving an unemployed person into a poor job 
match may actually worsen their mental health (Butterworth et al., 2011). 

Box 10 below contains more detail on methods used to calculate under and 

(50) However, life satisfaction does not worsen the longer a person is unemployed, although 
it remains lower than the satisfaction of employed people (see Clark (2006), Gordo (2009)).  

It is important to draw a distinction between self-reported job satisfaction and so-called 
‘objective measures’ of job quality. 

Studies of life satisfaction have not detected a sizeable negative association between 
part-time work or other poorer quality jobs and subjective job satisfaction (Layard, 2004).

In fact, some studies find a positive relationship, particularly for female workers (Bardasi 
and Francesconi 2004; Blanchflower and Oswald 1998; Booth and van Ours 2007; D’Addio 
et al. 2007; Manning and Petrongolo 2004; Wooden and Warren 2004)  

Box 10:Ways to measure over and under work and work/lifebalance

Data on life satisfaction confirms that people make different tradeoffs between work and 
leisure, and that life satisfaction is primarily affected by a mismatch between actual and 
preferred work/life balance.

Indeed, a 2009 Australian study using HILDA data found that longer hours themselves 
contributed to negative life satisfaction only when they were unwanted, but that when 
there was an hours mismatch, the impact was relatively large, about half the impact of 
becoming disabled and just under the negative wellbeing impact of being unemployed 
(Wooden et al. 2009). 

Similarly, large German studies have concluded that subjective measures of job quality
including job satisfaction are more influential on life satisfaction than so-called objective
measures (Grun et al. 2010, p. 305). 

For this reason, previous measures of work/life balance that assume uniform preferences 
across the population are relatively unhelpful. 

Jones and Klenow construct a welfare measure that explicitly includes the utility benefit 
from increased leisure time, as well as greater consumption, and lower levels of inequality 
and mortality. 

They calculate leisure time as the residual of the year after subtracting eight hours a day 
for sleep and the country’s average number of working hours per worker and multiply this 
by the ratio of employed workers to the full adult population (Jones and Klenow, 2011). 

However, this assumes that additional working hours are unwanted as they reduce leisure 

The Australia Institute’s GPI included a deduction for overwork.

It assumes that any change in average hours worked by full-time workers above the 1982 
level of 39.9 hours per week is involuntary, and values these additional hours worked at 
an average hourly wage rate. 

It seems unreasonable to assume that all of the change in average working hours is 

Studies that ask people how much of their additional work is unwanted have generated a 
wide range of answers. 

The Australian Work and Life Index suggests that 36 per cent of Australian employees 
experience overwork. 

Wooden et al. using HILDA data estimate that 25 per cent of employees are overworked, 
and a 2007 ABS survey of Employment Arrangements, Retirement and Superannuation 
found 21 per cent were overworked. 

We use the ABS figure and have built into the model an ability for the user to further dial 
this down.

The ABS Labour Force survey estimates about 7 per cent of the labour force experience 

We use data from these studies that ask people to nominate their preferred working hours 
to track the prevalence of job mismatch. 

There may also be scope to ask questions of users to generate data for this question. 

Given the importance of mismatched work hours rather than over or underemployment, 
both underemployment and overwork measures should be included. 

The weighting of these measures should be based on the relative impact of unemployment, 
underemployment and overwork on wellbeing.

Wooden (2009) suggests that underemployment has a relatively small impact on wellbeing
(-0.51 percentage points) compared to overwork (-1.58 percentage points) and unemployment 

c	Job satisfaction 

The association between job satisfaction and higher subjective wellbeing is well documented 
(Beutell, 2006; Tait et al., 1989). 

However, one careful longitudinal study fails to find a strong causal link between job 
satisfaction and wider wellbeing – suggesting either that life satisfaction tends to cause job 
satisfaction (including presumably greater attractiveness to good employers and/or better job 
selection) or that some common cause – such as disposition – drives both job and life 
satisfaction (Rode, 2004). 

Where our index has used the AUWI as a means of calibrating a number of sub-domains, 
the option to do so here is unavailable. 

In eschewing questions predicated on employment so as “to be applicable to all people”,(52)
the only relevant question in the AUWI asks how satisfied people are with what they are 
achieving in life, something that does not necessarily invite reflection on employment. 

The HILDA database could be more helpful in this regard, but we have not been able to find 
analysis on the HILDA database that permits us to check the relationship between job and 
life satisfaction. 

Further, it appears that the raw job satisfaction results in HILDA have barely moved over 
what is now a decade-long life, meaning that even if we posited that job satisfaction 
generated strong wellbeing effects, this would still have failed to produce any change in 
the index in the last decade (see Table 17 below). 

Personal e-mail correspondence with Prof Robert Cummins, Thursday 23rd June 2011. 

Table 17: Job satisfaction, 2001 to 2008 (means)

			2001 	2003 	2005 	2007 	2008 
Males: Satisfaction with: 
Total pay 		6.7 	6.8 	6.8 	6.9 	7.0 
Job security 		7.5 	7.8 	7.8 	8.1 	8.0 
Work itself 		7.6 	7.6 	7.6 	7.6 	7.6 
Hours of work 		7.0 	7.0 	7.1 	7.1 	7.2 
Work/life flexibility 		7.2 	7.3 	7.4 	7.4 	7.4 

Overall job satisfaction 	7.5 	7.6 	7.5 	7.6 	7.6 

Females: Satisfaction with: 
Total pay 		6.7 	6.7 	6.9 	7.0 	7.0 
Job security 		7.9 	8.0 	8.0 	8.1 	8.0 
Work itself 		7.7 	7.6 	7.6 	7.6 	7.7 
Hours of work 		7.3 	7.3 	7.3 	7.3 	7.3 
Work/life flexibility 		7.6 	7.6 	7.5 	7.6 	7.5 
Overall job satisfaction 	7.8 	7.8 	7.7 	7.7 	7.7 

Source: Wilkins et al., HILDA, 2011, p. 78 

However, just this year Butterworth et al. (2011, p. 6) reported the following finding: 

	While the difference in mean mental health scores of those in a job with one
	adverse condition and those in an optimal job would not be deemed clinically 
	relevant, the findings do indicate that, at a population level, relatively small 
	improvements in psychosocial job quality could yield widespread improvement 
	in the overall mental health of the Australian workforce. 

Accordingly, the area will be kept under review as the work with the HILDA database 
develops in case it provides an opportunity to improve the index. 

We will track changes in the rates of unemployment, underemployment and overwork, 
based on the data sources set out in the table below. 

Job related satisfaction domain of the HALE Index
		Indicator 			Narrative 

Unemployment 	Unemployment rate 	Unemployment fell from 2005 to 
		(trend) ABS Labour 	2008 before increasing during the 
		Force, Australia (cat. 	GFC. The wellbeing reduction 
		no. 6291.0) 		from unemployment is higher in 
					2010 due to population growth 

Under-		Underemployment 		The underemployment rate has 
employment 	rate (trend) ABS 		generally been increasing. The 
		Labour Force, 		wellbeing reduction from 
		Australia, Detailed 		underemployment is higher in 
		(cat. no. 			2010 due to both this and 
		6291.0.55.001) 		underlying population growth 

Overwork 	ABS Survey of 		Overwork rates have remained 
		Employment 		relatively constant. The wellbeing 
		Arrangements, 		reduction from overwork is higher 
		Retirement and 		in 2010 due to population growth 
		Superannuation (cat. 	
		no. 6361.0) 		

Job satisfaction 	n/a at this time 		n/a at this time 

12. 	Political and social capital
			a	Political capital
			b	Social capital

a	 Political capital

The OECD Better Life Index measures the quality of governance based on voter 

This produces the flattering result for Australia of being the world leader with voter
turnout of over 95 per cent. 

However, this is largely an artefact of compulsory voting and so is of little value as 
an indicator. 

The ABS MAP project includes a number of indicators of political engagement, 
including the proportion of female MPs and levels of informal voting. 

The latter measure may capture disaffection with government, and has increased 
over the last few elections, but only by a per cent or two. 

Further, we are unaware of any way either measure could be calibrated to wellbeing. 

We do have a more general indicator of Australians’ opinions about their democracy 
in the form of the Australian Unity Wellbeing Index, which asks respondents to rate 
their level of satisfaction with the country’s government. 

Unlike other aspects of the AUWI, this domain exhibits significant variability over time, 
especially in recent years. 

For the initial measurement of the HALE Index we have collected and recorded 
governance data. 

However, it seems clear that movements in the index do not predict broader 
self-assessments of wellbeing. 

Thus, we have given it a weighting of zero at this time. 

Should the index be provided in a form that permits users to provide their own weightings, 
we can update the series to allow them to do so. 

b	Social capital 

The situation for social capital is somewhat different. 

It seems clear that social capital is an important determinant of wellbeing. 

However, there are serious difficulties with including it in a summary index of wellbeing. 

As the SSF Commission observes (2008, 182): 

	[S]ocial connections bring benefits for health: as a risk factor for 
	premature death, social isolation rivals smoking (Berkman and Glass, 

	Evidence also suggests that social connections are powerful predictors of 
	(and probably causes of) subjective well-being. . . . 

	[S]everal (mainly US) studies suggest that both child welfare (infant mortality, 
	teen pregnancy, low birth-weight babies, teen drug use, etc.) and school 
	performance (drop-out rates, test scores) are robustly predicted by measures 
	of community social capital. 

However, despite convergence towards an agreed definition of social capital as “social 
networks and the associated norms of reciprocity and trustworthiness” (Ibid.), as the 
SSF Commission observes, “national statistics are still rudimentary”. 

The recently released OECD Better life index includes just one indicator of social capital 
or ‘community’, which is the proportion of people who feel they have friends or relatives 
to rely on in case of need. 

Australia is a leader in this regard, with 94.5 per cent of its inhabitants answering in the 
affirmative – putting us sixth among our OECD peers, behind Iceland, Ireland, 
New Zealand, Denmark and Sweden. 

There appears to be a clear correlation in this data between the size of a country and 
its performance (though it is far from uniform). 

Australia is the best performing country of its size in the sample. 

The very high numbers achieved by most developed economies also suggests that, like 
the internalisation of the health effects of pollution, to some extent wealthy societies 
have strong social capital, and that lack of social capital is more a source of low wellbeing 
among a relatively small minority. 

We have little firm evidence on which to base any weighting for this indicator. 

However, given the high percentage, even a relatively substantial weighting would leave 
changes in the index having little impact on the overall index through time. 

Largely because at any reasonable weight we would give the indicator, it is unlikely 
to have a substantial impact on the overall index, we use a weighting of zero. 

However as with the measure of satisfaction with our political system, we can 
update the series over time in the event that the index is provided in a form that 
enables people to calculate the index according to weights different to the ones 
we have chosen. 

As the field matures it may well be appropriate to provide more expansive coverage 
of social capital and with it enable this sub-domain to have more impact on the overall 

For instance Helliwell and others (Helliwell and Huang, 2008; Helliwell and Barrington-Leigh, 
2010) argue that levels of community identification and trust can powerfully improve 
self-evaluated wellbeing both generally and in specific circumstances such as in 

Again, however, it may be that Australia already enjoys high levels of trust and that, 
unless that trust deteriorates substantially, which seems unlikely, little impact 
will be had on a general measure of wellbeing. 

Issue 		Indicator 					Preferred Weight (%) 

Confidence in 	AUWI Satisfaction with Government Index 	Zero 
Social capital 	% of people who have someone to rely on in 	Zero 
		time of need (Gallup World Poll) 

Home | Index |O'view |Glossary | Part 1 | Part 2 | Part 3 | Part 4 | Part 5 | References | Appendix on Method 
	 Acknowledgements | Boxes - 10 | Figures - 11 | Tables - 18 | Disclaimer | Copyright

THD F 2012/08/17 10:00 - 12:17 Rendering content for Web-page THD F 2012/08/17 13:15 - 16:32 Preparing Web-page THD F 2012/08/17 18:00 - 19:05 Checking Web-page for launch on Internet