Introduction:
The focus of this paper is on macroeconomic management and not on the entirety of economic policy. There are many issues which macroeconomic interventions cannot address. To use macroeconomic instruments, rather than the relevant targeted instrument, will blunt the effectiveness of macropolicy interventions.
Reflecting, this paper is really a critique of the current primary macropolicy indicator – GDP. Economists are too much creatures of habit for one to expect it to be replaced. At least we need to remember its limitations. While it dominates, macropolicy will have little to do with wellbeing.
Evaluating Wellbeing
There are two major problems concerning wellbeing that this paper is dealing with. The first is what determines an individual’s wellbeing.
Once upon a time economists equated wellbeing with ‘utility’, but the discipline has moved past that. (One methodological problem was that utility could not be defined independently of the theory which it is being used to justify.)
Instead, thinking has moved towards focusing on answers to subjective questions such as ‘are you happy?’ or ‘rate yourself on life satisfaction’. There is evidence that an individual’s objective physiology tends to be consistent with subjective responses, although it is not conclusive (and probably cant ever be).
Numerous studies have found consistent patterns with the subjective responses, most notably that age, employment status, gender, health, marital status and (even) ethnicity all seem more important.
(Note that it has been difficult to test/measure some obvious potential influences on wellbeing. For instance one’s wellbeing is likely to be higher if the wellbeing of close associates (such as immediate family) is also. It also seems likely that one’s wellbeing may be influenced by that of the society as a whole – an implicit idea behind the ‘team of five million’.)
People with higher incomes than others report higher wellbeing (measured, say, by happiness) but not by much. (Even here, it is possible that the effect is from being higher in the rankings of a society which measures status by income, which has quite different policy implications from if the wellbeing lift comes from the actual goods that the income can purchase.)
However (and surprisingly) raising across-the-board incomes in an affluent economy does not raise across-the-board wellbeing. (In contrast, in traditional analysis .economic growth was expected to raise wellbeing. It does, however, raise choice, especially in respects to Sen’s notion of ‘capabilities’, which also undermines traditional economic analysis.)
Historically economists assumed that higher income resulted in greater wellbeing. However, that no longer seems as true in affluent economies.
Economic policy has little impact on age, employment status, gender, health, marital status or ethnicity, which all seem more important for determining wellbeing. However, the evidence is that employment status (such as being stressfully unemployed) is more important than income.
The research does hint that there are significant gains from income increases for those on the lowest incomes, but the effect is not well measured. (There are likely to be other longer-term gains in health and for children’s prospects.)
In summary then, income is only marginally relevant in an affluent economy, yet the main paradigms have hardly been modified by it. This is as true as aggregate wellbeing, which might be the objective of macroeconomic policy interventions.
That raises the second issue. How to aggregate individual wellbeing for aggregate policy, such as macropolicy.
With the above background we consider how wellbeing might be incorporated into macropolicy, knowing the exercise is difficult, but is not as treacherous as relying solely on GDP.
(It could be added that economics is in even greater flux. The subject is older than psychology as a science and had to adopt simplistic assumptions about behaviour. However psychology has been testing those assumptions and found many are inconsistent with actual behaviour. Again, the findings have not been really incorporated into the standard paradigm, although at least six Nobel laureates in economics have contributed to the revolution. As far as I can judge, the findings do not generally impact on macroeconomics although their impact on microeconomics is likely to be great.
Below I add Maslow’s hierarchy of needs to the analysis.)
Summary: The Key Propositions
1. The way the game is scored shapes the way the game is played. (Gilling’s Law)
2. This paper is about macroeconomic policy – fiscal, monetary, and exchange rate interventions (alternatively it may be defined as aggregate demand management).
3. The easy way to lose a game is having multiple scoring systems.
4. Menus and pantries are checklists not ultimate indexes of performance.
5. GDP per capita is not a very good measure of wellbeing.
6. Income is no longer a good measure of wellbeing for most people.
7. GDP and related aggregates are not equity-neutral but favour the rich. A better measure is Real National Income (RNI).
8. Government spending needs to be included at a proper value to the recipients rather than cost.
9. The calculation of the RNI should include government services provided to the individual
10. Employment Status is a part of Wellbeing.
11. There is no evidence that price stability or inflation impacts directly on wellbeing.
12. The Primary Macropolicy Wellbeing Indicator will not contain much about sustainability.
Proposition 1: This paper is framed by Gilling’s Law: The way the game is scored shapes the way the game is played.
Proposition 2: Different games have different scoring rules. This paper is about macroeconomic policy – fiscal, monetary, and exchange rate interventions. (An older view, or perhaps a newer view, is that it is about aggregate demand management.) It is not, for instance, about the concerns of a productivity and wellbeing commission.
Proposition 3: The easy way to lose a game is having multiple scoring systems. Faced with them, there will be a default to a particular dimension. In the case of macropolicy this is likely to be GDP which, as we shall see, is not a very good measure of wellbeing. (Proposition 5). Thus the objective here is to identify one scoring regime (a single objective), although supplementary assessments are briefly explored. (Proposition 12)
It is argued that perhaps there should be a number of indicators with policy aiming to locate the economy within set bounds for each one. (E.g. consumer inflation between 1 and 3 percent p.a.) I am not uncomfortable with that approach (assuming the bounds are sensibly wide enough, and consistently obtainable). However the reality is that while the macroeconomic managers may have little difficulty with that approach, the public discussion will fixate on one particular indicator (GDP?). That will influence the politicians and in turn compromise the macroeconomic management. (An example of this sort of fixation is the net-public-debt-to-GDP ratio – more at Proposition 12.)
Proposition 4: There may be a role for menus (or pantries) but as checklists not as objectives. For instance, the Treasury has a wellbeing framework of five dimensions, useful for a Treasury official who, when finishing a paper, checks whether Treasury’s main concerns have been covered.
(My objection to the five dimensions is that there should be a sixth one of ‘quality’. It is an extraordinary omission given that many economic interventions are about quality. I assume that the oversight occurred because the Treasury adopted a limited neoclassical paradigm which says quality is not a problem in a market economy.)
Proposition 5: GDP per capita is not a very good measure of wellbeing.
It measures output not income (which differs if there are changes in the terms of trade and border transaction costs).
It measures gross output gross including capital depreciation.
It measures output in a jurisdiction, not the income of those in a jurisdiction. (Another ambiguity is to what extent temporary visitors and recent arrivals should be included. It is possible that all the output gains in the Key-English decade accrued to them.)
Because of these considerations, a better national accounting measure would be the SNA measure of National Income.
(Some of the other weaknesses of GDP, such as its treatment of externalities, non-market activity and resource depletion, are not so relevant for macropolicy. See however, Proposition 12 in regard to the last.)
Proposition 6: Income is no longer a good measure of wellbeing.
This is a recent research finding which has surprised the profession. Historically economists assumed that higher income resulted in greater wellbeing. However, as discussed in the introduction, that seems no longer as true as it once was.
What is happening may best be explained in terms of the Maslow hierarchy of needs (from bottom to top: physiological needs => safety needs => love and belonging needs => esteem needs => self-actualisation needs.) The economy’s primary contribution is to the bottom one (physiological needs). It is not quite true that the economy does nothing towards the higher one (for instance economic stabilisation may contribute to safety and security needs; possessions to status-esteem needs, but generally the role of conventional economics measures such as income become very less relevant in the way the research reports.
(Some economic measures may impact badly on the higher needs. For instance, pursuing greater economic output may decrease employment stability and reduce the ability to obtain safety needs.)
Proposition 7: GDP and related aggregates are not equity neutral but favour the rich.
When economics tried to avoid interpersonal comparisons it shifted to a Pareto criterion that an increase in output was good, because the output could be allocated as a distribution authority saw fit. However practically, we operate on the principle that any redistribution will reduce output (See Proposition 8). Therefore redistribution is discouraged.
Amartya Sen asked why we assume that a dollar to a rich person has the same social value as a dollar to a poor person. (That is an assumption which has been sneaked in, despite the alleged avoidance of interpersonal comparisons.) He suggested that a better assumption would be to treat a one percent increase in income to a rich person the same as a one percent increase in income to a poor person and proposed a measure which he called ‘real national income’ (RNI).
RNI is difficult to calculate given the current data base but one can use the standard household equivalent income calculations from the HYES data (i.e. not the national accounts). Between 1982 and 2018, real private incomes based on the GDP/Pareto measure rose 1.2 percent p.a. but because of the rise in inequality (incomes rose faster for the rich than the poor) the annual rise on a RNI/Sen basis was 1.0 percent. The cumulative difference was 6.0 percent. In effect, the RNI level was in 2018 where the GDP level was the same in 2013/5; growth had been retard by a month a year as a result of distributional policy.
While I have taken a Sennian approach here, at the distributional policy level the Atkinson approach (the Atkinson index, also known as the Atkinson measure or Atkinson inequality measure). Proposition 8 is relevant to it. (For an extension, see Proposition 9)
Proposition 8: Government spending needs to be included at full value rather than cost.
Government services are measured in the SNA framework at cost, whereas private production is measured in terms of the value to the consumer. We know something about the value of government services – or at least the Treasury thinks it does.
It takes that $100 of government spending (as measured conventionally) displaces $112 of private production. The logic which follows is that the government commits itself to spending only what it values the service at 112 percent of the nominal cost.
It is not hard to see why this might happen. For instance, the provision of public healthcare is considerably cheaper than private provision via, say, insurance. The additional cost of private insurance is thought to be more expensive than (12/112 =) 11 percent of total costs (which is a reason why US health care is so expensive but relatively ineffective).
Moreover, when some New Zealander receives $100 of public healthcare the other five million are likely to be pleased that the expenditure has happened because it gives them some comfort that if we need the service they will get it too. (Were they each to pay .01 cent for that comfort, the value of the $100 of care would be $150 to the nation.)
To add a premium of 12 percent to government services to convert from production costs to value would not necessarily change our international GDP ranking, since the same adjustment would be made for other countries’ GDP (although their premium might be different and the proportion of aggregate production devoted to government services also varies).
However, it would change the the growth of GDP when there was a change in the private to public expenditure balance. For instance, the share of public consumption (excluding public investment) was 19.7 percent in the 2009 year but only 18.0 percent in the 2017 year. Allowing for value rather than cost, the effect of the Key-English Government squeeze was to reduce effective GDP by 0.2 percent.
(One may be doubtful about the 12 percent premium, but it is the Treasury estimate. In the context of the above analysis, those on the left may think the premium was too low and those on the right think it too high.)
(As an aside – or is it? – the Atkinson approach has been around for about 50 years, but I have never seen it applied in New Zealand. (I have never had the resources to do this; others have.) Given its relationship to Rawls’s Theory of Justice which has also been around for fifty years and also has had little impact on New Zealand social policy, perhaps one is not surprised.)
Proposition 9: The calculation of the RNI should include government services provided to individuals (e.g. health and education – social transfers are already including in the RNI calculation).
They contribute to a person’s wellbeing Any increase or reduction of such services should be monitored.
I am not sure we currently have a database to do this on an annual basis.
(This adjustment is likely to reduce RNI growth relative to GDP growth even further, especially when big public spending cuts were made in the early 1990s.)
Proposition 10: Employment Status is a part of Wellbeing
A standard research finding is that wellbeing is reduced by stressful unemployment (which is a different notion from the conventional unemployment measure and probably includes dire underemployment).
There is a linkage here to Maslow’s hierarchy together with Jahoda’s latent functions of work. (Not in Narrow Seas: 410-1) The central insight is that work does not just provide income but has other personal and social benefits.
The logic of the research is that there is a tradeoff between higher output and higher unemployment. For instance, before 1984 there was a conscious policy of job creation to reduce unemployment even if it reduced productivity. After 1984 that policy tradeoff was abandoned, with the prioritisation of high output/productivity (the goal may not have been achieved) even though that raised the level of (stressful) unemployment. Wellbeing was diminished.
The tradeoff needs to be incorporated in the PMWI. Left among subsidiary indicators, as the unemployment rate is today, relegates its significance and means the primary indicator is not a comprehensive measure of economic wellbeing.(Note the stress from lower income is covered by a RNI measure; the concern here is the loss of the Jahoda/Maslow benefits.)
Proposition 11: There is no evidence that price stability or inflation impacts directly on wellbeing. While the inflation rate is a relevant concern for macroeconomic policy as a part of ensuring the market economy works properly it is not a part of a PMWI.
(In passing, the ‘Misery Index’ of adding together the unemployment rate with the annual inflation rate is an obvious mathematical nonsense – since the components are measured in different units – it also lacks conceptual foundations.)
Proposition 12: The Primary Macropolicy Wellbeing Indicator will not contain much about sustainability.
The PMWI is a measure at a point in time, the current state of wellbeing. Trying to incorporate sustainability over time will destroy any meaning. Instead, two supplementary indicators are recommended.
There is a need for a measure of financial sustainability. Currently the government debt to GDP ratio is used. Its defects are obvious. (Credit rating agencies look at much wider sets of indicators.) Financial sustainability also needs to include the overseas financial position and private balance sheets. To progress this to something meaningful requires a national balance sheet.
There is also a need for an indicator of environmental sustainability, especially as reducing carbon emissions impacts on macroeconomic policy. This is quite different from the (usually uninformed) attempts to extend GDP to cover the environment. (The trick, by the way, which determines their outcome is how the environment is valued. If the value is high relative to output, then the indicator goes down, if it is low the indicator rises – funny that. A related trick is to fiddle with the time-discount rate.)
Conclusion
The paper points out the humbling (for an economist) proposition that many of the greatest influences on wellbeing may no longer be particularly influenced by economic management.
Instead it explores the development of a Primary Macropolicy Wellbeing Indicator. The focus has been on the construction of a single indicator because multiple indicators will be misused – typically by defaulting to an inappropriate one. But because a macropolicy indicator is at a point in time there is also a need for two supplementary indicators to assess sustainability to add a time dimension.
In particular the paper explores the replacement of the current primary indicator – GDP – by something which takes better into account key elements in our understanding of wellbeing. Thus the paper reviews standard economic criticisms of GDP/output as a measure of social wellbeing.
Any alternative has to be able to be measured and able to be forecast to be useful. The paper does not propose a detailed PMWI but suggests how one might be constructed.
Among the issues the analysis has raised are
– net income rather than gross output (Proposition 5)
– what is the relevant population? (Proposition 5)
– distributional considerations (Proposition 7)
– how to value wellbeing enhancing government services (Proposition 8 )
– unemployment (Proposition 10)
– sustainability (Proposition 12)
By playing down each of these and giving primacy to the GDP as the macropolicy indicator results in a bias against wellbeing in outcome. Arguably that has been a feature since GDP became the dominant indicator, say, four decades ago and as a result wellbeing has suffered.
In particular, there are many instances of when ignoring these wellbeing effects in favour of GDP has damaged wellbeing. There are may instances during the period of Rogernomics/Neoliberal dominance but there are also examples from the receny Key-English period.
The rough estimates in the paper suggest that moving towards a PMWI makes significant but not great differences to the current measure. What is important it changes the way we think about the impact of macroeconomic policy on wellbeing.
I do not expect New Zealand policy to replace GDP with something more suitable in the near future. Our policy framework is conservative, both in the sense of intellectual inertia and in favouring a pro-rich status quo. In any case the current framework is colonially subservient to overseas fashion. However, we might make a little progress – pay a little more attention to wellbeing – if we keep the critique made here of GDP more in the forefront of our thinking.