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


1. Introduction and background

2. What is required (Called "Our Goal" in the text)
	a	Intellectually rigorous and comprehensive 
	b	Commonsensical
	c	Responsive to emerging developments
	d	Engaging
	e	Politically and ideologically bipartisan

3. The choices available
	a	‘Correcting’ GDP measures
	b	Subjective wellbeing measures
	c	Composite indices
	d	User weighting of indices
	e	‘Dashboard’ approaches

1. Introduction and background

Since at least the ‘marginal revolution’ in economics in the 1870s, 
debate has raged as to how best to conceptualise economic wellbeing, 
though the debate on human wellbeing goes back to at least the 
beginning of philosophy. 

Ironically, as economists debated whether it was legitimate to compare 
wellbeing or utility among individuals or whether to adopt the more 
fastidious Pareto criterion of welfare,(3) an alternative means of 
measuring wellbeing arose from the emerging practice of national 

Today gross domestic product, or GDP, is routinely taken as a 
touchstone of economic progress. 

It is also taken to be a measure of economic wellbeing sometimes if 
not explicitly then implicitly, by politicians, economic commentators 
and the media. 

Yet measures of GDP were conceptualised and built as a measure of 
economic activity rather than wellbeing, and it is well understood 
that for that reason they can be misleading. 

To name a few obvious foibles, GDP captures production exchanged 
within the market and so abstracts from domestic production. 

Thus parents making and selling sandwiches at the school tuckshop 
contribute to GDP while they do not if they make the same sandwiches
at home. 

More starkly, sexual activity does not contribute to GDP – unless 
it is prostitution. 

GDP also measures national production, not consumption, which one 
might argue is the whole point of production. 

Crime harms society and individuals, but at least in the short term 
it can add to GDP as the destruction it wreaks is not registered in GDP, 
yet the investment to rebuild damaged property and to guard against 
further crimes – through more police on the beat and investment in 
security technology – does contribute to GDP. 

Similarly, poor health significantly reduces wellbeing, but the impact 
on GDP is ambiguous.

If it keeps people from working this lowers GDP, but if the cost of 
clinical intervention outweighs the loss of wages working then at 
least in the short term GDP rises. 

And the running down of our natural or social capital does not register
in GDP but may have significant impacts on both short and longer-term 
levels of wellbeing. 


The Pareto criterion holds that one can be confident that one has brought about a social improvement in welfare only 
if everyone in the new state of affairs is as well off as they were, while some person or people are better off. 

The criterion can be useful in theory, but owning to the complexity of the world, it is rarely useful in practice, 
for very few changes lead to winners without there being any losers. 

Time spent on leisure pursuits is considered inherently unproductive 
within the GDP framework, although for some it is the most valuable 
time to a person (and the ultimate end of working). 

In fact these debates were alive as national accounting was being
established in the middle part of the 20th century and were certainly 
well established as a point of complaint by the 1970s. 

They were a major theme of Hugh Stretton’s Boyer lectures (1974) and 
of Fred Hirsh’s The Social Limits to Growth (1976), which pointed 
to the increasing extent to which consumption becomes ‘positional’ as 
income rose. 

In the early ’70s Norhaus and Tobin (1972) proposed a series of 
rearrangements to items in the National Accounts to create what they 
called a primitive and experimental measure of economic welfare.(4)

At the same time the Easterlin paradox was documented – beyond a 
certain relatively modest point in economic development the effect of 
further increases in incomes on increasing reported happiness 
encounters severely diminishing returns. 

The debate has broadened further since then. 

During the 2000s, economist Richard Layard revisited Easterlin’s paradox
with contemporary research arguing that, above a certain level of income,
happiness does not correlate particularly well with it. 

Another criticism of national accounting measures of wellbeing is that 
not just GDP but also the kinds of corrected national accounting 
measures suggested by Norhaus and Tobin give a materialistic bias 
to the measurement of wellbeing. 

Australia is not alone in revisiting these issues. 

Globally, interest in better measures of wellbeing is increasing. 

Recently the Stigliiz-Sen-Fitoussi (SSF) Commission has explored 
this terrain and made a range of proposals designed to expand 
the focus of statistical indicators from economic production 
to broader measures of wellbeing and sustainabilty. 

Composite wellbeing indices are being constructed in Canada 
and the UK. 

While this report was in preparation, the OECD released its Better 
Life Index, which allows readers to compare quality of life across 
countries based on their own personal rankings of the relative 
importance of different aspects of life.(5) 

Much activity is also occurring at the institutional level. 

The UN has recommitted to finalise an international statistical 
standard for the production of 


Their proposed new ‘Measure of Economic Welfare’ [MEW] was constructed by removing from GDP components
that are capital in nature (such as the replacement of obsolete assets) or instrumental goods that are made
in order to limit the impact of harm rather than create wellbeing. 

Government expenditure on defense and policing are included in this category. 


OECD Better Life Index

Imputed valuations are also made for leisure time and non-market work. 

The UN has recommitted to finalise an international statistical standard 
for the production of a set of Satellite National Environmental Accounts. 

The Federal Treasury has developed its own approach to economic wellbeing 
around five principles. 

	(i) 	Level of opportunity and freedom that people enjoy 
	(ii) 	Level of consumption possibilities 
	(iii) 	Distribution of consumption possibilities 
	(iv) 	Level of risk that people are required to bear 
	(v) 	Level of complexity in people’s lives. 

There have been numerous attempts to correct for the inadequacies of GDP 
as a measure of wellbeing, although a consensus approach has not yet 

As we outline in the sections below, five main approaches have been 
attempted – corrected GDP measures, measures of subjective wellbeing, 
composite indices, user weighted and dashboard approaches. 

The Herald/Age -Lateral Economics (HALE) Index of Wellbeing seeks what 
is best from each of these approaches. 

2. Our Goal

The aim of the project has been to build a wellbeing index that might be 
updated regularly. 

Such an index should satisfy the following criteria to the maximum extent 

It should be: 

a	Intellectually rigorous and comprehensive

	or as rigorous and comprehensive as such an exercise can reasonably 
	aspire to be. 

b	Commonsensical. 

	Much ink has been fruitlessly spilled in pursuit of intellectual 
	rigour and of building ‘value free’ foundations for conceptualising 
	human wellbeing. 

	Yet the desire to boil down measurements over a range of dimensions 
	makes such a quest unhelpful. 

	Ultimately the intellectual tools used must be matched to the purposes 
	at hand as commonsensically as possible. 

c	Responsive to emerging developments. 

	One challenge for such an indicator is that most social and economic
	change happens very slowly with volatility only at the margin. 

	Further, where one is taking subjective measures of wellbeing, 
	aggregate subjectively reported wellbeing is often relatively stable 
	over large groups and, as the Easterlin paradox suggests, changes 
	little once a certain income level has been achieved.

	This creates a challenge to capture changing experience in a 
	meaningful way, which will reflect changes experienced over months 
	rather than decades. 

d	Engaging. 

	It should interest, intrigue, stimulate and satisfy the reader. 

	As a measure of national wellbeing it should be accessible to 
	a wide audience. 

e	Politically and ideologically bipartisan.

	It should be seen as a ‘fair go’ at a difficult problem rather 
	than a tendentious exercise in rehearsing its own, or its readers’,

3. The choices available 

It is possible to discern five possible approaches to developing the 
HALE Index of Wellbeing. 

a	‘Correcting’ GDP measures 

The accounting framework that underpins the national accounts does not 
only generate values for GDP. 

It is possible to use the quarterly national accounts to construct other 
metrics of economic progress. 

In particular, rather than focusing on the value of goods and services 
produced in Australia, we can look instead at the value of national income 
this production creates, for it is income that 
ultimately supports higher rates of consumption and living standards. 

Box 1:GDP and NNI

Gross Domestic Product is often used as a measure of economic performance, 
but it has three major drawbacks in this respect. 
It’s Gross – that is, depreciation of physical and natural capital is not 
It’s Domestic – that is, it measures output produced in Australia, even 
though the resulting income may flow overseas 
It’s a Product – the ultimate aim of economic activity is not production 
in itself but the income it generates, which should be taken to include 
the economic value of leisure, household work and so on..... 
But, if we want to look at policies that promote our economic welfare in 
the long term, we need to start with another measure, produced by the same 
National Accounts that give us GDP, but with the errors above fixed. 

That measure is Net National Income (NNI): the amount of income accruing 
to Australians, after replacing depreciated capital 

Source: John Quiggin, Blog post, May 6th, 2010.(6) 

Indeed the Stiglitz-Sen-Fitoussi (SSF) Commission’s very first recommendation 
was to shift towards measuring income rather than production. 

It found measures of Net National Disposable Income are the most comprehensive 
available (Stiglitz et al., 2009, p. 93ff). 

These matters are explored further in the next section. 


The Central Flaw in the Heanry Review 

However, even if we recalibrate GDP to transform it from a measure of 
economic activity within the market (which reflects the origins of national 
accounting in its attempt to systematise the smoothing of the cycle of economic 
activity) into an analogous measure of the consumption possibilities to which 
such activity gives rise, this is far from a comprehensive measure of welfare.
Indeed NNI remains far from a comprehensive measure of economic welfare, 
let alone welfare more generally. 

Like GDP it measures only the market sector of the economy, so the impact on 
wellbeing of time available for leisure or time spent in non-market activities 
such as caring for children are not accounted for.
It also does not account for changes in non-physical capital such as the 
discovery or depletion of natural resources or the generation and atrophy of 
human capital. 

If we can do it adequately, correcting for these things would lead to a 
comprehensive measure of economic wellbeing. 

However, it would still fail to illuminate other aspects of life that 
most of us regard as ingredients of our wellbeing such as our physical 
and mental health and the health of our environment. 

b	Subjective Wellbeing Measures [SWB]

A critical concept for the architects of the ‘marginal revolution’ in economics 
from the 1870s through till the turn of the 20th century was the concept of the 
‘utility’ of various goods and services to consumers. 

Competitive markets would equilibrate prices and marginal costs and consumers 
would purchase goods and services up to the point at which price equalled the
‘marginal utility’ that specific goods or services might provide them. 

Of course ‘utility’ was never observable directly, but, like the ether in 
19th-century physics, it was a metaphysical construct that seemed to be 
implied by the framework that was being adopted. 

Utility proved to be a mixed blessing for the new approach and led to numerous 
controversies, including the question of whether one could make legitimate 
interpersonal comparisons of utility. 

Leading English economists Marshall and Pigou argued that there would be 
diminishing marginal utility of income, which meant that, other things being 
equal, social utility was increased if a dollar of income was moved (for 
instance via taxation) from a rich person to a poor one. 

For the dollar would go from meeting discretionary or even luxury needs to 
meeting urgent ones. 

The Italian philosopher and economist Pareto challenged interpersonal 
comparisons of utility and proposed the criterion of welfare improvement, 
which found its way into neoclassical economics. 

One can think of the new and burgeoning field of subjective wellbeing (SWB)
as a revisiting of the spirit of the early marginalists – an attempt to put 
flesh on the metaphysical bones of ‘utility’ by asking people about their 
subjective wellbeing. 

As we argue below, this may not be the killer move that its proponents might 
have hoped for, but the SWB literature contains important information that
can assist in building a useful, convincing and engaging index of wellbeing. 

Box 2:Subjective Wellbeing Measures

The Australian Unity Wellbeing Index (AUWI) provides a six-monthly reading
of Australians’ subjective judgement of their own personal wellbeing and 
satisfaction with national wellbeing. 

Participants are asked to rank their satisfaction with various aspects of 
life and society (see Table 3: AUWI - Aspects of Life Measured) against a 
scale where 0 is completely dissatisfied and 10 is completely satisfied. 

The AUWI is reported as a satisfaction percentage. 

Table 3: Australian Unity Wellbeing Index-Aspects of life measured 

Personal Wellbeing 		National Wellbeing 

1. standard of living 		1. economic situation 
2. health 			2. state of the environment
3. achievements in life 	3. social conditions
4. personal relationships 	4. how Australia is governed 
5. how safe you feel 		5. business 
6. community connectedness 	6. national security 
7. future security 
8. spirituality/religion 

The Household, Income and Labour Dynamics in Australia (HILDA) survey has 
also included questions regarding life satisfaction since its inception in 2002. 

Unfortunately, subjective wellbeing measures have important limitations. 

Though asking people what they think about their own wellbeing makes 
obvious sense, the answers still fail to engineer a clean transmission from 
subjective reporting to objective fact. 

If two people rate themselves 7 out of 10 in SWB, can we really conclude that 
they enjoy equivalent wellbeing? 

Leaving aside basic questions of honesty, one subject may be calibrating their
reporting of their own wellbeing against a stoical standard, while another 
subject reports against a self-indulgent one. 

This may reflect any number of factors from personal idiosyncrasies of factors 
related to age, gender or culture. 

This and the different nuances of words in different languages make cross-country 
comparisons of SWB subject to an additional difficulty. 

General measures of wellbeing tend to hover around 75 out of 100 for most 
people and over large numbers are very stable and so a poor indicator of 
changes in satisfaction. 

Significant short-term changes over 1 percentage point to reported happiness 
levels have occurred only four times in the life of the Australian Unity 
Wellbeing Index, with no clear trend discernable. 

On the other hand, measures of subjective wellbeing (SWB) can provide useful 
information with which to cross validate other data and to provide a 
methodology for comparing different aspects of life and wellbeing. 

For instance, researchers can use SWB analysis to ask questions like, 

“What amount of money would increase average life satisfaction by one
percentage point?” 

If some life event – say, a bereavement – tends to reduce life satisfaction 
by a similar amount, one can then plausibly claim some equivalence between 
the two events. 

We use this methodology to roughly calibrate the relative value of an 
additional increment of income in improving the life satisfaction of 
wealthy and less wealthy people. 

c	Composite indices

The other approach commonly used is to combine various measures of 
economic wellbeing into a single composite index.

This approach need not start with the National Accounts. 

Indeed, many attempts note the limitations of GDP as a measure of overall 
wellbeing, and instead collate a large number of other indicators thought 
to represent dimensions of wellbeing. 

This approach has its attractions, particularly the ability to present a 
richer array of data and avoid the need to place monetary values on 
non-economic aspects of wellbeing, so they can be compared to GDP. 

However, composite indices still require value judgements about which 
indicators to include and how to weight them in constructing an overall 

While value judgements ultimately cannot be avoided in this area, as the 
Stiglitz-Sen Commission complained, the authors of existing composite 
indices seldom made these normative implications explicit or put forward 
a rationale justifying their decisions (Stiglitz et al. 2009, p. 65). 

Canada’s now fully operational Index of Wellbeing (CIW) and the Kingdom 
of Bhutan’s famous measure of Gross National Happiness (GNH) are both 
composite indices. 

Other attempts to present a richer, more multi-faceted exploration of wellbeing 
include the UN’s Human Development Index (HDI).
Rather than attempting to adjust GDP figures themselves to account for 
unvalued or misvalued elements of economic wellbeing, the HDI creates a new 
indicator from weighting existing measures of health (life expectancy) and 
education levels (mean years of schooling). 

Both the HDI and CIW consolidate indicators to generate a single composite 

This approach has not been adopted in Australia to date, although the 
Australian National Development Index [ANDI] project led by Professor Mike 
Savaris at RMIT is developing such an approach. 

Box3: The Tyranny of Equality

‘Splitting the difference’ has obvious appeal as an anchor in bargaining 
between two people or two perspectives. 

Something similar often occurs when weighting various possibly incommensurable
components of wellbeing – whether in putting together the final composite 
index or in assembling sub-indices measuring different domains such as 
‘environment’ and ‘social’ wellbeing. 

Thus for instance the Kingdom of Bhutan’s composite index of GNH is quite 
sophisticated in its methodology. 

Yet at each turn in their construction of the index, its architects weigh 
each measure equally. 

Ostensibly this has been done “in order to avoid bias”.(7)

But should ‘time use’ (Dimension 2) receive equal weight to education or 
‘standard of living’ (Dimensions 6 and 8)? 

Would not extreme poverty or illiteracy be a worse fate than a bad time 
imbalance in one’s life? 

Weighting dimensions equally is just as much a choice as is differential 
weightings, though the latter is likely a sign of choice by design rather 
than default. 

Further, if weightings are equal, the introduction of an additional measure
that is then given equal weighting then downgrades the significance of the 
original measures. 

Thus, the number of sub-indices that comprise each of the nine domains of 
Bhutan’s GNH ranges from three to 11 sub-domains. 

It is often difficult to understand why these various dimensions are counted 
equally within sub-domains but unequally within the global GNH. 

Even in indices that assign uneven weights to different indicators, a closer 
look suggests these are in fact just a variant on the ‘split the difference’

For example, the Yale Environmental Performance Index (EPI) includes 25 
performance indicators across 10 policy domains, which have a relative weight 
of between 0.694 per cent and 25 per cent of the total index. 

However, though they have tried to reflect the consensus of environmental 
scientists and policymakers their weightings are more art than science. 

In fact, these weightings really reflect a nested cascade of ‘splitting the 
difference’ decisions. 

Thus impacts on humans and nature are weighted 50:50, then impacts on nature 
are split 50:50 between climate change and other natural resources, and then
within these sub-domains the primary indicator usually receives a 50 per cent 
weighting compared with other indicators being equally weighted 
(See Table 16 on page 54) 


GNH - 9 Domains 

d	User weighting of indices

Some indices allow those using them to recalibrate the weightings they apply. 

The Australia Institute’s GPI’s website (now discontinued) allowed users to vary 
its own preferred weightings on a website. 

Likewise the OECD’s Better Life Index puts the weightings entirely in the hands 
of the visitor to its website, although when the visitor arrives for the first 
time on the index’s website the weightings between sub-domains are equal which 
must operate for some as a default, whether this was intended by the website’s 
designers or otherwise. 

‘Dashboard’ approaches 

Another approach is to accept the essential incommensurability of different 
aspects of wellbeing and to report them without attempting to encompass them 
within a single summary index. 

In some ways this is the most intellectually respectable and certainly the 
most intellectually safe method. 

Where Canadian and UK Governments have moved towards single indices of 
wellbeing that make such valuations and trade-offs explicitly, Australia’s 
Bureau of Statistics has been a leader in the development of satellite welfare 
measures to augment the national accounts (Salvaris, 2009, p. 2). 

Its Measures of Australia’s Progress (MAP) reduce various aspects of these 
welfare measures into indices over specific domains – for instance the health 
of inland waters would measure the health of many inland waterways and 
aggregate the results in a single measure of progress or decline. 

However it does not aggregate its reporting beyond this level and instead 
offers a ‘dashboard’ that displays whether we are progressing, standing 
still or regressing in a range of areas under three general themes: 

	A	Society, 

	B	Economy and 

	C	Environment. 

The ABS has explicitly chosen not to attempt to consolidate information 
into a single composite index. 

A similar compendium and dashboard-based approach was the Blair/Brown UK 
Government’s Sustainable Development Indicators [SDI] project. 

The SDI includes 68 indicators based on 126 underlying measures, with a 
focus on the direction of rather than the magnitude of any positive or 
negative change over time (UK Department of Energy, Food and Rural Affairs, 

In 2010, MAP showed that where headline indicators were available, social and 
economic indicators had generally improved over the last decade, but that 
environmental indicators had deteriorated (see Figure 2). 

Figure2:Measures of Australia's Progress [MAP]

Many of the progress indicators used in MAP are directly comparable to the 
indicators chosen by the CIW in creating the wellbeing domains that underpin 
its composite index.

Accordingly, if the methodology underpinning the CIW is robust, we can, in 
principle, transform the ABS MAP data in a similar way. 

Figure 3 below shows a simple translation of MAP’s headline indicators (where 
available) into a composite index, based on the methodology of the CIW.(8) 

Data for each progress indicator is transformed into an index, where 
the 1999-2000 is the base observation equal to 100. 

For years where no data is available, the value of the index is interpolated 
as a straight line average of known values. 
A composite index for each of the economy, society and environment 
domains is created by averaging the indices for each domain’s 

All indicators receive equal weighting. 
The overall composite progress indicator is then just a simple average 
of the economy, society and environment indices. 


That is, each sub-domain indicator is equally weighted to create an index. 

Figure3:Simple MAP Index,2000 - 2008
Simple Index from ABS Measures of Australia's Progress Headline Indicators 

As the ABS has not yet settled on headline indicators for a number of sub-
domains, a MAP Index including only domains with a headline indicator would 
not be as comprehensive an index as the CIW. 

It would be open to us to choose our own indicators for areas in which MAP 
has yet to settle on a headline indicator. 

As Table 4 shows, comparable Australian data already exists for almost all 
the indicators used by CIW. 

However, there are two concerns with such an approach: 

*	It is not clear that a greater number of indicators delivers a more 
	accurate picture of wellbeing. 

	Averaging across a large number of indicators reduces the impact of 
	changes in particular variables, making the overall index less likely 
	to change over time. 

	Equally, the Canadian methodology assumes all indicators are equally 
	important for wellbeing. 

*	Australian data for many of the indicators would be quite old (often 
	based on data collected in 2006). 

Table4:Comparison of indicators used in CIW and Australian data

Other measures of wellbeing in Australia tend to follow the dashboard 
approach including the Community Indicators Project in Victoria and numerous 
dashboard or ‘triple bottom line’ style reporting initiatives being undertaken 
at the local government level. 

It is clear that a dashboard approach is the least likely to attract legitimate 

In order to produce a single, quantitative index, one assumes the commensurability 
of different aspects of human experience and wellbeing and that requires heroic 
assumptions to be made. 

This is not to mention a more fundamental problem, which is that a single index
of wellbeing necessarily aggregates all people’s wellbeing when each person would 
weight the importance of different things very differently (if they could weight 
them at all!) 

On the other hand, it can be a worthwhile discipline to attach weights to 
different aspects of experience given that policy decisions must constantly be 
made that make tradeoffs between those dimensions of wellbeing at the 

For instance, if we improve the health of our rivers or our population 
and this consumes more economic resources than it generates, then we should 
do so only if the wellbeing dividend exceeds its economic cost. 

But we cannot know this without some summary index of community wellbeing. 

Further, while it is important that the index be as intellectually rigorous as 
possible, rigour is only one of many criteria that must be jointly optimised. 

The index must also engage and educate, and making weightings explicit to be a 
better way of encouraging community debate about such things than simply 
leaving it to individuals’ personal preference. 

For what it is worth, in our weighting of the criteria, we claim no more than our 
own values and our own common sense. 

Others will disagree, and still others will insist that collapsing the dimensions 
of wellbeing into a single index remains a folly. 

These are reasonable views. 

Yet ultimately we cannot agree. 

Amatya Sen, whose work on capabilities forms the theoretical backbone of the 
UN HDI, is himself wary of summarising the wealth of data into a single index. 

Yet he relented in his view, having been persuaded that only a single index 
could shift policy-makers’ attention from material output to human wellbeing as 
a real measure of progress (Fukuda-Parr, 2003, p. 305). 

His judgement appears correct in hindsight with the HDI being highly influential 
in steering development policy towards a broader definition of welfare than is 
dreamt of in the philosophy of national accounting. 

In the following section we set out the way in which we can use national 
accounting and other data to correct GDP of some of its most glaring flaws. 

We then move on to trying to address some more fundamental problems with GDP 
in measuring economic wellbeing – particularly its neglect of changes in the 
capital stock. 

This then forms the basis for a more satisfactory core of economic welfare to 
which later sections then add measures of other aspects of welfare that cannot 
reasonably be captured in national accounting. 

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