Download as pdf or txt
Download as pdf or txt
You are on page 1of 31

doi: 10.1111/j.1467-6419.2010.00635.

HAPPINESS AND HEALTH: TWO PARADOXES


Simone Borghesi and Alessandro Vercelli
University of Siena
Abstract. This paper aims to establish systematic relationships between the two rapidly growing
research streams on the socio-economic determinants of happiness and health. Although they have
been pursued quite independently by different communities of researchers, empirical evidence
points to very similar underlying causal mechanisms. In particular, in both cases per capita income
seems to play a major role only up to a very low threshold, beyond which relative income and
other relational factors become crucial for happiness and health.
On the basis of these structural analogies, we argue that a process of cross-fertilisation between
these two research streams would contribute to their development by clarifying the relationship
between happiness, health and their determinants. Finally, we observe that the two literatures have
converging policy implications: measures meant to reduce poverty and inequality and invest in
social and environmental capital may improve both health and happiness of the individuals.
Keywords. Happiness; Health; Happiness paradox; Poverty; Inequality; Relational goods.

1. Introduction
The causal interaction between happiness and health is well documented. Empirical evidence shows
that self-reported happiness of people closely depends on their perceived health status, and at the same
time the happiness of individuals deeply affects their health status. People who feel happy enjoy a
better health, while unhappiness deteriorates their health status reducing the immune resistance and
originating psychosomatic diseases that may lead to depression and suicide.
The socio-economic determinants of happiness and health have been extensively studied in the
last decades, although almost always separately, by different communities of researchers: mainly
by economists, psychologists and sociologists in the case of happiness, principally by social
epidemiologists in the case of health. We intend to show that the analysis of the links between
these two rapidly growing streams of empirical analysis may be illuminating for both of them helping
us to get a better understanding of the relationship between happiness, health, their determinants and
their policy implications.
The structure of the paper is as follows. Section 2 explains the reason why in this work we chose to
focus on the interaction between subjective indexes of happiness and objective indexes of health. Section
3 examines the main determinants of subjective happiness in the light of the recent empirical evidence
accumulated by economists, psychologists and sociologists. Section 4 considers the main determinants
of objective health in the light of the recent empirical contributions provided by social epidemiologists.
Section 5 discusses the link between happiness and health and the common mechanisms underlying
their relationship with income. Section 6 briefly comments on the policy relevance of the results
emerging from the comparative analysis of the two research streams on happiness and health. Section
7 concludes.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233



C 2010 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden,

MA 02148, USA.
204 BORGHESI AND VERCELLI

2. Indexes of Happiness and Health


We can define a whole spectrum of measures of happiness that ranges between two extremes: ‘objective’
measures based on the direct observation of physiological states, and ‘subjective’ measures based on
their direct assessment (see Frey and Stutzer, 2002). At one extreme, we have subjective measures
based on self-reported evaluations. These measures are at the centre of the literature on happiness. At
the other extreme, we have objective measures in the strict sense of the word based on the controlled
observation of asymmetric brain waves (Frey and Stutzer, 2002).
Although the terms ‘objective’ and ‘subjective’ are commonly used in the literature that we are
going to revise, the distinction between these two categories is fairly ambiguous since also the alleged
‘objective’ measures of happiness obviously present a high degree of subjectivity. In fact, scholars in
this field generally label as ‘objective’ the measures of happiness deriving from the direct observation
of empirical evidence, but in the case of happiness the latter heavily depends on the subjective status
of the person being observed.
We then have a grey zone in which sophisticated methods are used to correct the subjective distortions
of self-reported measures. In this sense, the resulting happiness indicators are also sometimes called
‘objective’. A particularly rigorous attempt of this kind is that pursued by Kahneman and collaborators
who start from instantaneous subjective assessments, considered more reliable than assessments of
remembered experiences, and devise integration procedures of the instantaneous measures that reduce
the subjective biases of remembered happiness (Kahneman, 1999). This research stream made progress
in the last years but is far from being completed. In this work we ignore both classes of so-called
‘objective’ measures of happiness as their methods are still under scrutiny and the empirical evidence
produced is still insufficient for our comparative purposes.
Similarly to the literature on happiness, health indexes are also labelled as ‘subjective’ or ‘objective’,
depending on whether they are based on self-reported assessment or on the direct observation
of empirical events such as death and diagnosed illnesses. Also in the health literature, however,
the distinction between ‘objective’ and ‘subjective’ health measures appears fairly arbitrary. Life
expectancy, for instance, cannot be directly observed in any country at any point in time, but it
is computed through models that entail subjective judgements about the future mortality rates, the
probability of catastrophic events that may increase such rates (e.g. wars or epidemics) and about how
to deal with missing observations.1
Another important category of ‘objective’ health indexes that has been recently developed is given
by quality-adjusted health indexes that integrate mortality and morbidity indexes with other indexes
of life quality to combine the quantity and quality of an individual’s existence. In particular, the
World Health Organization (WHO) provides estimates of Healthy Life Expectancy (HALE), namely,
the number of years that a person can expect to live in ‘full health’ (WHO, 2004). This measure, that
is obtained by subtracting the numbers of years of ill-health from the overall life expectancy, can shed
light on the effective life quality of the individuals. In the Organisation for Economic Cooperation and
Development (OECD) countries, for instance, HALE estimates are about 9–10% below life expectancy
and it turns out that the burden of ill-health is higher for women than for men, so that although women
live longer than men (80.6 versus 74.7 years) they also experience 2 more years of ill health during
their lives (OECD, 2005). The HALE estimates certainly represent a significant step forward in the
direction of health indexes that may capture important aspects of individual happiness. However, much
remains to be done. As a matter of fact, HALE estimates currently present several problems in terms of
reliability and comparability between different countries due to the different methodologies being used
(OECD, 2005). Moreover, the notion of ‘full health’ is very difficult to define as well as the weight to
be given to different ill-health status deriving from physical pain or mental disease.2
Objective health indexes are no doubt partial indexes of the health status that is effectively an
unobservable variable. However, they certainly provide a better measure of the health status than

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 205

Figure 1. Life Satisfaction and Per Capita GDP in 2005 (140 Countries).
Source: Authors’ elaboration on Abdallah et al. (2009).

self-reported judgements, which are strongly conditioned by the personality, and culture of the self-
assessing individuals. They are able, moreover, to capture the impact of a few factors that are remarkably
important for health. In particular, they implicitly take account of unconscious sources of happiness
that by definition do not emerge in subjective assessments.
In what follows, therefore, we will focus mainly on the determinants of stated happiness and of
predicted life expectancy, that represents the most general and best-known ‘objective’ measure of the
health status of the population, and on the links between these two indicators.

3. The Determinants of Subjective Well-Being: The ‘Happiness Paradox’


In this section, we aim to discuss the main determinants of subjective happiness in the light of the
empirical evidence accumulated in the last decades as a benchmark for the comparison with the socio-
economic determinants of health and the study of their reciprocal interactions and policy implications.

3.1 The Role of Absolute Income and Income Aspirations


Until recently, per capita income has long been considered as the main determinant of individual
happiness. This deep-seated conviction has been only partially corroborated by recent empirical
research. Cross-section empirical analysis generally detects a positive correlation between subjective
happiness and per capita income both across countries and within countries. Figure 1 shows the
relationship between these variables that emerges in a sample of 140 countries using data on per capita
Gross Domestic Product (GDP) in the year 2005 and life satisfaction as reported by Abdallah et al.
(2009).3
As the figure shows, an increase in per capita income Y appears to increase significantly subjective
happiness W ∗ only for very low classes of income; the impact of absolute income on happiness tends
to diminish for higher income classes and to fade away after a relatively low threshold estimated to
be at around $10,000 per year (according to Frey and Stutzer, 2002) or $15,000 (according to Layard,
2006)
W ∗ = f 1 (Y ), f 1 > 0, f 1 < 0 (1)
Moreover, if we analyse the evolution of reported subjective happiness in developed countries after
World War II (WWII), the correlation with per capita income is generally nonexistent or slightly
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
206 BORGHESI AND VERCELLI

negative in apparent contradiction with the predictions of standard utility theory. This result was first
found by Easterlin (1974) for the United States and subsequently confirmed for other developed
countries such as the UK, Japan, France, Germany and the Netherlands (Easterlin, 2001). This
documented decoupling between the empirical trends of income and self-reported happiness is often
referred to as the ‘happiness paradox’. In fact, an increase in personal income extends the set of goods
that may be purchased and consumed and this is supposed to improve the total utility (or happiness) of
rational agents. The paradox becomes even more surprising if we accept the deep-seated conviction of
traditional utilitarian economic theory that the relevant measure of welfare is consumption not income,
because consumption propensity progressively increased in developed countries in the last decades.
Several theories have been set forth to explain the happiness paradox. One of the main explanations
is that, as the psychologists pointed out long ago, any differential of hedonic experience (pleasure
or pain) induced by a specific event is short-lived and rapidly fades away. Any attempt at increasing
the happiness of individuals appears thus condemned to defeat; following Brickman and Campbell
(1971), this worrying effect is often called ‘hedonic treadmill’ (see also Bottan and Perez-Truglia,
2008; Graham, 2008). This is true in particular of economic consumption. The utilitarian tradition
was aware of this psychological law and focused mainly on non-durable consumption whose transient
effect on happiness is obvious. Following Scitovsky (1976) we have to distinguish, however, between
comfort goods whose contribution to happiness is particularly short-lived and stimulation goods that
have a longer, and possibly persistent, impact on happiness as their consumption stimulates individuals’
creativity. Although the confessor of Henry IV complained for having been served for too long with
his favoured dish (partridge), a fan of Beethoven would not complain for having the opportunity of
listening again to his masterpieces as each performance may reveal new delights. In the case of durable
goods, the distinction between comfort and stimulation goods is even more crucial. The rapid decline
of pleasure given by the consumption of comfort goods leads to the frustrating attempt at progressively
upgrading the comfort contents of the goods consumed, so triggering a form of consumerist addiction
(Scitovsky, 1976). On the contrary, stimulation goods may keep alive an increase in happiness much
longer. While buying the last model of television or car brings about a short-lived differential of
happiness, the purchase of a guitar or a book may have a durable impact on it. The hedonic treadmill
contributes to the explanation of the happiness paradox but is insufficient to fully account for it because
an increase in income should translate in durable happiness to the extent that consumption focuses on
stimulation goods.
The ‘theory of livability’ stresses the role of a deep-seated needs hierarchy because income growth
satisfies the primary needs of individuals but not by itself the superior needs related to their self-
realisation (Veenhoven, 1984). This theory captures a further crucial factor that explains why income
growth has a diminishing effect on happiness. As soon as basic needs are satisfied, happiness cannot
be kept alive simply by increasing consumption of superfluous and luxury goods but only by shifting
to superior goods that enhance the self-realisation of individuals. Only by pursuing the latter course,
happiness may improve in a sustainable way.
The theoretical explanations reviewed so far are not necessarily inconsistent with traditional utility
theory.4 The causal factors already examined are insufficient, however, to fully explain the happiness
paradox. Other explanations have thus been put forward that cannot be easily reconciled with traditional
utilitarianism unless we are prepared to modify and extend it in a substantial way (for a recent attempt
of this kind, see Clark et al., 2006).
According to the ‘aspiration theory’, suggested by Easterlin in his seminal paper (Easterlin, 1974),
the happiness of economic agents depends not on the outcomes of their behaviour alone, in particular
their per capita income, but on the gap between aspirations of income Y ∗ and effective income Y

W ∗ = f 2 (Y ∗ − Y ), f 2 < 0 (2)
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 207

Because desired income is considered to be a positive function of effective income, the gap is
continuously reproduced (‘satisfaction treadmill’). Whatever is the explanation of the dynamics of
income aspirations, the latter tend to grow with effective income according to most empirical estimations
(cf. Frey and Stutzer, 2002)
Y ∗ = k(Y ), k≈1 (3)
∗ ∗
From this empirical regularity we easily derive that W = f 2 (Y − Y ) ≈ 0, that is that
happiness does not tend to increase in consequence of an increase in per capita income.
A variant of this theory emphasises the role of positional goods in explaining the continuous shift
of aspirations (Hirsch, 1976; Frank, 1985). The conspicuous consumption of positional goods can
be interpreted as a zero-sum game: any attempt at keeping one’s social position increases the agent’s
income by reducing her leisure, therefore it induces more intense personal efforts without affecting one’s
satisfaction. This variant is strictly connected with the role of social factors and relative income since
a high consumption level may be seen as the scarcity rent of socio-economic status (see next section).
Because the latter is in fixed supply, any conspicuous consumption imposes negative externalities on
other individuals. The aspiration theory in all its variants may thus account for the so-called ‘satisfaction
treadmill’ but cannot explain by itself the progressive decline of subjective happiness observed in many
developed countries after WWII unless we assume a growing frustration of aspirations that remains
groundless (Ng, 1978). Moreover, the elasticity of aspirations to effective outcomes is unlikely to
maintain a value close to one along all the life cycle of individuals. In particular, it seems plausible to
assume that aged people have a progressively lower elasticity of aspirations to income.
In the last few years, the happiness paradox has been the object of renewed interest that has stimulated
an exponential growth of the literature in different research fields (mainly economics, psychology and
sociology) meant to explain, qualify or deny this apparent paradox. The evolution of the debate has
been heavily influenced by the progressive growth of the available data sets that are increasingly broad
and self-consistent. In his pioneering contributions, Easterlin had to work with very limited, rough and
often heterogeneous data referring to less than 10 countries. A significant step forward came from
the World Values Survey that progressively extended the size of the sample, from 20 countries in the
first wave (1981–1984) up to 69 countries in the fourth wave (1995–2004). In 2002, the Pew Global
Attitudes Survey interviewed 38,000 respondents in 44 countries. A further important advance came
quite recently from the 2006 Gallup World Poll that surveys subjective well-being across 132 countries.
On the basis of an accurate analysis of the new data sets, several authors (e.g. Clark et al., 2008; Di
Tella and MacCulloch, 2008; Deaton, 2008; Stevenson and Wolfers, 2008; Easterlin, 2009; Easterlin
and Angelescu, 2009; Easterlin and Sawangfa, 2009) have recently reviewed the happiness–income
relationship, which has led to a reinterpretation of the previous evidence in the light of the recent
one. Although most of the recent contributions seem to support the existence of a happiness paradox,
there is not unanimous consensus in the literature. Stevenson and Wolfers (2008), for instance, express
serious doubts on previous findings and question the very existence of the happiness-income paradox:
‘our key result is that the estimated subjective well-being-income gradient is not only significant but
also remarkably robust across countries, within countries, and over time’ (Stevenson and Wolfers, 2008,
p. 3). More precisely, using the 2006 Gallup World Poll data, they find that in cross-section analysis
the gradient of happiness does not diminish with per capita income. Contrary to preceding findings,
moreover, they detect a positive correlation between happiness and per capita income also in the time
series analysis performed on Japan and Europe (but not on the United States).
These results – that have been interpreted as strong support to the negationist view – might depend,
however, on the confusion between short- and long-term correlation of the variables. As Easterlin and
Angelescu (2009) have pointed out, the time-span of the short-run regressions conducted by Stevenson
and Wolfers (2008) is too brief (between 5 and 6 years) to identify the long-run relationship between
happiness and per capita income. In particular, for some of the transition countries included in their
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
208 BORGHESI AND VERCELLI

database, the time-span covers only the recovery phase of the transition (in which life satisfaction
grows along with per capita income), but is too short to capture both the collapse and the recovery
phases, as it would be needed to make proper inferences on the long-term trend. In fact, if the analysis
performed by Stevenson and Wolfers (2008) is repeated omitting the transition countries, the slope
coefficient of the happiness–income relationship turns out to be statistically insignificant.
We may conclude that the new data allowed Stevenson and Wolfers and other authors (see, in
particular, Deaton, 2008) to better circumscribe the paradox but in its updated version it is still present.
As data progressively improved and empirical evidence accumulated, the focus of the debate shifted
from the lack of correlation between happiness and per capita income in general, to the lack of
correlation in time series analysis, and finally in long-run time series. Although the puzzling aspects
have been progressively circumscribed la pièce de résistance that remains unscathed calls for a deeper
understanding of the additional factors that may affect happiness beyond per capita income in the long
run.

3.2 The Role of Relative Income and Social Factors


As we have seen, after a threshold between $10,000 and $15,000, the influence of absolute income on
happiness rapidly fades away. Recent research suggests that, after this threshold, relative income Y R
and social factors are likely to replace absolute income as the major driving force affecting happiness.
As a matter of fact, happiness generally decreases if the relative personal income of individual i
diminishes relatively to the (average) per capita income of a reference group j and vice versa. The
simplest formalisation is the following:
Wi∗ = f 4 (Y R ), f 4 > 0 (4)
where Y R = Yi /Yj and Yj is the (average) per capita income of a reference group (the Joneses,
colleagues, people with similar education or a similar job, and so on). This correlation holds also
for more sophisticated measures of income inequality such as Gini or Theil indexes. Graham and
Felton (2005), for instance, find that in Latin American cities self-reported well-being is largely
affected by both objective measures and self-reported perceptions of inequality, particularly for the
lack of social mobility that the latter implies. Drawing upon opinion data from the European Social
Survey, also Biancotti and D’Alessio (2007) find that inequality tends to lower subjective well-being
in Europe: people dislike inequality either for the conflicts, tension and social unrest that it may cause,
or for the perceived unfairness of an unequal income distribution, depending on personal inclinations
and views. Alesina et al. (2004) examine how cultural backgrounds and personal values affect the
individual perception of inequality in America as compared to the European Union. They show that
the relationship between happiness and inequality varies across the two continents according to the
way inequality is interpreted by individuals: the American poor generally do not dislike inequality as
they perceive it as a sign of social mobility, whereas the happiness of the European poor is negatively
affected by inequality because they perceive it as indicating little chances for their future improvement
along the social ladder.
Although the consensus is not unanimous in the literature (see O’Connell, 2004; Haller and Hadler,
2006 for notable exceptions), many other studies (e.g. Hagerty, 2000; Fahey and Smyth, 2004) have
found that income inequality is negatively related to happiness. Since in most OECD countries income
inequality slightly decreased up to the late 1970s and then increased, this factor cannot explain by
itself the behaviour of self-reported happiness since WWII. However, it may contribute to explain the
worsening of the trend observed in many developed countries since the early 1980s.
Another interesting explanation of the happiness paradox stresses the role of relational goods R
W ∗ = f 5 (R), f 5 > 0 (5)
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 209

The category of relational goods has been introduced only recently in economic theory in order to
capture the affective and communicative components of interpersonal relations, the importance of which
has emerged in economics in different fields (Nussbaum, 1986; Gui, 1996; Antoci et al., 2009). These
goods have quite different characteristics from those of ordinary goods as they are end in themselves,
cannot be produced or consumed by a single individual but only simultaneously by at least two of
them, and their value depends on the interaction between individuals under conditions of reciprocity
(for a recent assessment, see Gui and Sugden, 2005). Examples are love, friendship, and more in
general direct personal social relations, that is not mediated by economic or political exchanges.5
Many empirical studies find that relational goods play a crucial role in the determination of self-
reported happiness. Intense and frequent social interaction is positively correlated with happiness while
loneliness is negatively correlated with it (Deci and Ryan, 1985). More generally, intimate relationships
that are perceived as stable and safe (e.g. having regular sex with the same partner, being married or
an unmarried cohabitor since long, seeing family and friends on a regular basis) tend to have a positive
effect on individual happiness, while the ending of such relationships (e.g. personal or parental divorce,
widowhood, and so on) turn out to be harmful for the well-being (e.g. Heliwell, 2003; Blanchflower
and Oswald, 2004a; Lelkes, 2006; Pichler, 2006; Dolan et al., 2008). Bruni and Stanca (2008) on the
basis of World Values Survey (around 264,000 observations coming from more than 80 countries for
the period 1980–2001) showed that there is a strong positive correlation between subjective happiness
and the length of time dedicated to relational activity (in the family, with friends, in voluntary service).
Relational goods may have the nature of stimulation goods as they do not necessarily contribute to
the comfort of individuals but appeal to their superior needs. The contribution to happiness given by
social interaction is higher when it is not motivated by self-interest. For example, empirical evidence
shows that altruism and voluntary service contribute to happiness (Frey and Stutzer, 2002; Bruni and
Stanca, 2008). According to many researchers, the progressive development of the market has displaced
and suffocated the process of production and consumption of relational goods and this contributes to
explaining the happiness paradox (Bartolini and Bonatti, 2002). One economic reason underlying this
process is a continuous alteration of the relative price of comfort goods that progressively decreased due
to technical progress and standardisation, while the costs of relational goods did not diminish or even
increased (Bruni, 2006). This led to a process of substitution of potentially fulfilling but demanding
relational goods with cheap comfort goods. A case in point is the increasing time dedicated to television
watching as a surrogate of relational activity: empirical evidence shows that TV watching is positively
correlated with hours of work and negatively correlated with happiness (Bruni and Stanca, 2008). The
deterioration of relational goods in the last decades contributes to explain the happiness paradox, but
its slow trend suggests that this factor is unlikely to explain by itself the happiness paradox.
Two other important factors that have a crucial social dimension are unemployment and education.
It is widely accepted that unemployment reduces well-being, even after controlling for the associated
fall in income (Clark et al., 2006)
W ∗ = f 6 (U ), f 6 < 0 (6)
Clark and Oswald (1994, p. 655) conclude their accurate research on the effects of unemployment
in the UK asserting that ‘joblessness depressed well-being more than any other single characteristic’.
These results have been confirmed for many other countries (an early survey of the literature may be
found in Darity and Goldsmith, 1996, and articulated updating in Frey and Stutzer, 2002 and Dolan et
al., 2008). The unemployed is victim of anxiety, anger and depression and suffers from a loss of self-
esteem and social status that may disrupt his/her family life and social life. One may wonder whether
unhappy people may possibly select into unemployment for being less productive, having poorer health
or personal choice of quitting their jobs, so that the direction of causality between unemployment and
unhappiness could be reversed. However, empirical studies consistently find that selection effects are
absent or minimal (Winkelmann and Winkelmann, 1998; Lucas et al., 2004), even after controlling for
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
210 BORGHESI AND VERCELLI

psychological distress before unemployment (Korpi, 1997) and for individual heterogeneity using fixed
effects models (Ferrer-i-Carbonell and Gowdy, 2007). Unemployment, however, is not strictly correlated
with the evolution of happiness in many countries. While the latter stagnated in many industrialised
countries in the post-war period, in the same countries unemployment diminished in the 1950s and
1960s, increased in the 1970s and 1980s, and diminished again in the 1990s and first 2000s. The latter
reduction, however, was correlated to growing flexibility of labour markets and industrial relations that
is likely to have contributed to declining job satisfaction. The strong correlation of unemployment with
unhappiness seems inconsistent with the conviction that unemployment is never involuntary (see, e.g.
Lucas, 1981, and, for a criticism, Vercelli, 1991; Frey and Stutzer, 2002). This implies that policies
meant to eliminate involuntary unemployment, in the absence of negative collateral effects, could
contribute to increase the happiness of workers.
Another important determinant of happiness is the degree of education I

W ∗ = f 7 (I ), f 7 > 0 (7)

Along the whole life cycle a higher degree of education correlates with a higher degree of subjective
happiness and the differential is independent of variations of per capita income (Blanchflower and
Oswald, 2004b; Easterlin, 2005). Moreover, as it emerges from some empirical studies on specific
countries (Bukenya et al., 2003 for the United States; Gerdtham and Johannesson, 2001 for Sweden),
higher education turns out to have also a relevant indirect effect on happiness through its positive impact
on health. Education increases the set of enjoyable goods as it expands, for example, the fruition of
cultural goods (literature, classical music and opera, theatre, cinema, and so on). The higher the level of
education, the higher is the capability of appreciating creative and stimulating goods and activities that
are much less subject to the hedonic and satisfaction treadmills. In addition, the upgrading of the average
education level would contribute to the diffusion of well-informed preferences reducing unnecessary
unhappiness (Scitovsky, 1976; Easterlin, 2005). The average level of education increased continuously
in most countries since WWII; however, apparently, this did not contribute to an improvement of
happiness, probably because, as in the case of income aspirations, individual aspirations to higher
education increased continuously with the average education level, thus determining a ‘satisfaction
treadmill’ also for education.

3.3 Other Factors


The ‘theory of adaptation’ maintains that individuals are characterised by a stable equilibrium state of
happiness. When this state is perturbed by a positive or negative event the effects are only temporary
as the individuals rapidly adapt to the new conditions (Brickman et al., 1978)

Wi∗ = f 8 (Wi∗ − Ŝi ), f 8 > 0 when Wi∗ − Ŝi < 0, and f 8 < 0 when Wi∗ − Ŝi > 0
(8)

where Ŝi is the stable equilibrium point (or ‘set point’) of happiness of individual i. As for the
determination of the equilibrium point, a few psychologists maintained that individuals have their own
level of happiness that is independent of their experience and is firmly rooted in their own personality
as established by genetic and psychogenetic factors G (the so-called ‘theory of personality’ or ‘set
point theory’; see, e.g. Likken and Tellegen, 1996)

Wi∗ = f 9 (G i ) (9)

Although the importance of personality and genetic factors on happiness is well established, the
strong version of this theory has been recently questioned by accurate empirical studies that show that
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 211

the personality of individuals can only partially explain the variations of their self-reported happiness,
and that in any case adaptation is not complete (Diener, 1996).
Another influential theory stresses the fact that the growth of per capita income produces negative
externalities E that deteriorate the happiness of the individuals
W ∗ = f 10 (E), 
f 10 <0 (10)
An extensive literature documented huge environmental negative externalities E in the post-war
period (see, e.g. Borghesi and Vercelli, 2008, and the literature therein cited). Among them, a crucial
role was played by the external diseconomies produced by pollution and exhaustion of environmental
goods that contributed to the depletion and deterioration of the environmental capital.6 In addition, we
have to consider the deterioration of social capital produced by inequality and the growing influence of
modern mass media (in particular television). This factor provides a crucial explanation of the happiness
paradox but the quantification of its impact is strictly dependent on the list of relevant externalities and
their evaluation, issues that are still very controversial.
Finally, we have to mention the health H of individuals as a major determinant of their happiness.
Frey and Stutzer (2002, p. 56) remark that ‘when people are asked to evaluate the importance of various
areas of their lives, good health obtains the higher rating’, so that
W ∗ = f 11 (H ), 
f 11 >0 (11)
Because the impact of health on happiness is particularly important and quite complex we postpone
its analysis to Section 5.

3.4 Concluding Remarks


The determinants of happiness that we have considered so far do not exclude each other. Each of them
captures an important causal factor of self-reported happiness. Although there is a serious problem
of multi-collinearity, for each of the factors considered above there are several empirical studies that
argue in favour of their, at least partial, independence. We can thus summarise the main acquisitions
of the research surveyed above through the following function of self-reported happiness
W ∗ = F(Y , Y − Y ∗ , Y R , R, U , I , G, E, H ) (12)
where the partial derivatives with respect to each variable are assumed to have the same sign of the
derivatives discussed earlier. Each of the capital letters that appear in the argument of the function F
may be considered as a vector of variables, some of which have the dimension of flow and others
of stock. Thus, for instance, the happiness of people depends on their per capita income but also on
their wealth, on their aspirations about both income and wealth, on flow externalities but also on stock
externalities concerning the quantitative and qualitative characteristics of environmental and social
capital, on flow exogenous variables (such as genetic shocks) and stock exogenous variables (such as
the genetic endowment of each individual), on education and human capital, employment status and
curriculum.
Of course, other variables may be added among the determinants of the happiness function (12).
However, their role has been insufficiently explored and their independence inadequately supported
by the empirical evidence. Possible exceptions could be leisure, inflation and institutional factors.
Traditional utility theory contrasted work-generating disutility with leisure generating utility (Argyle,
1996), but this is not true in general. An interesting work may contribute to happiness (Frey and Stutzer,
2002, and the literature therein cited) while not all leisure activities contribute to happiness: bowling
alone or watching television may be a sign of solitude and depression (Putnam, 2000; Bruni and
Stanca, 2008). The necessary disaggregation of the leisure activities sends back, therefore, to factors
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
212 BORGHESI AND VERCELLI

that have been already considered above such as social relations, absolute and relative income, health.
Another important candidate to be included between the main happiness determinants is inflation.
Frey and Stutzer (2002) found a negative correlation between inflation and happiness, whereas Di
Tella et al. (2001) estimated the happiness trade-off between unemployment and inflation. One may
doubt, however, whether inflation has a genuine independent influence on happiness. Inflation acts
mainly through modifications of absolute and relative income and disruption of social relations, which
sends back to the factors already considered above. Finally, a few authors found a clear correlation
between happiness and a few institutional factors (see, in particular, Frey and Stutzer, 2002). This line
of research is inspiring but it is not yet clear to what extent the influence of institutions on happiness is
independent of their impact on factors already considered above such as social relations and relational
goods.
We may thus interpret the function (12) as a fairly good representation of the main factors affecting
self-reported happiness. This relation explains pretty well the happiness paradox. We may agree that
different individuals are characterised by different propensities to happiness descending from their
personality and rooted in their genetic patrimony. The amazing adaptive capabilities of human beings
assure a gradual convergence after shocks towards their characteristic happiness equilibrium. The latter
cannot be conceived, however, as a stationary equilibrium since the positive and negative shocks are
not completely reabsorbed in a reasonable time spell. Therefore, the outcomes of individual behaviours
(e.g. per capita income) may have a persistent impact on happiness, but this depends on the gap between
outcomes and aspirations since the latter tend to shift in the same direction of the outcomes. It is still
difficult to explain the substantial lack of correlation between income and happiness in rich countries. A
further contribution comes from the theory of needs hierarchy that introduces an evolutionary element
in the analysis. We are still unable to explicate, however, why free time did not increase in the last
decades. This may be accounted for by the role of positional goods and short-termism that increased
the indebtedness of families and by the changes in labour markets that jeopardised the stability of jobs
and the scope of workers rights.
In more abstract terms, income and wealth growth tend to expand the economic liberty of the
individuals (and therefore, in principle, also their happiness) as they enhance the set of available
economic options. The trouble is that in recent times this process came along with increasing constraints
to the production and consumption of goods that reduce the economic liberty of the individuals and their
happiness so that the net effect is uncertain and depends on specific structural and policy circumstances.
Summing up, the paradox of happiness is not at all a paradox. We are not lacking for an explanation;
on the contrary, we have perhaps an excess of explanations. They are all related to factors neglected
or under-emphasised by GDP statistics. We have to conclude that the real paradox is the persisting use
of GDP statistics as the crucial index of the individuals’ well-being.

4. The Socio-Economic Determinants of Health


In this section, we intend to analyse the main socio-economic health determinants in the light of
the recent advances in social epidemiology. This will allow us to get a deeper understanding of the
relationship between happiness and health, and of the common features that characterise the two
research streams.

4.1 The Role of Absolute Income


The per capita income of a community (at a local, national or international level) is generally considered
as a major determinant of its average health. This causal link is confirmed by an extensive set of
empirical studies focusing on both cross-section and time series analysis. An increase in per capita
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 213

Figure 2. Life Expectancy and Per Capita GDP in 2005 (140 Countries).
Source: Authors’ elaboration on United Nations data (UNDP, 2007).

income relaxes the budget constraints imposed by severe poverty that hinder the capacity of a person
to prevent and cure a disease. In addition, an increase in average per capita income in a country is
generally accompanied by higher expenditure in health programmes, by better medical infrastructures,
and by updated medical knowledge and know-how. For these reasons, economic growth is generally
regarded as the main driving force for the remarkable increase of average life expectancy at birth that
has been observed at the world level in the last decades (from 47 years in 1950–1955 to 65 years in
2000–2005 according to United Nations, 2005). The empirical literature shows that the pattern of the
relationship between absolute income Y and health H is very similar to that of the relationship between
absolute income and happiness discussed in Section 3.1
H = ϕ1 (Y ), ϕ1 > 0, ϕ1 < O (13)
In cross-section studies it emerges that the health of the poor has a much higher income elasticity
than that of the rich. In particular, cross-country evidence suggests that life expectancy increases with
average per capita income in relatively poor countries, whereas this relationship tends to vanish for
relatively rich countries (Preston, 1975). This can be clearly seen by looking at Figure 2 that shows
the relationship between life expectancy and per capita GDP in the year 2005 based on United Nations
data (UNDP, 2007) referring to 140 countries.
Similar results emerge also in single-country cross-section studies. Using a survey on health and
income in Britain, Wilkinson (1992) finds that several health indicators increase rapidly as income rises
from the lowest to the middle classes of income distribution, whereas no further health improvements
occur at higher income levels. Similarly, using data from the National Longitudinal Mortality Survey in
the United States, Deaton (2002) observes that the male (age adjusted) probability of death decreases
rapidly as income grows at low family income levels whereas it flattens out at high family income
levels.

4.2 The Role of Relative Income and Social Factors


In recent years, several studies have found that socio-economic inequality is positively and significantly
related to the individuals’ health, particularly in developed countries (Borghesi and Vercelli, 2004).
Several contributions in different disciplinary fields have found that, after a threshold of minimum
income (at about $5,000), income inequality emerges as a crucial independent determinant of health
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
214 BORGHESI AND VERCELLI

even when controlling for other factors including absolute income (Wilkinson, 1992). Similar results
emerged in several other studies that focused on different groups of countries and periods of time (see,
e.g. Cantarero et al., 2005; De Vogli et al., 2005; Leigh and Jencks, 2007).
The same relationship was found also at the local level. Comparative analysis across 50 US states,
for example, showed a close relationship between inequality and mortality rates (Kaplan et al., 1996).7
Analogously, among the 282 US metropolitan areas, the ones with the most unequal income distribution
turned out to have the highest mortality rates (Lynch et al., 1998). Similarly, De Vogli et al. (2005)
have found that among Italian regions income inequality has had an independent and more powerful
effect on life expectancy at birth than per capita income and educational attainment.
Although these regressions did not control for some further explanatory variables and there is not
yet unanimous consensus in the literature on the evidence available,8 these and similar results obtained
in other countries seem to suggest that relative income, independently of absolute income, can have a
crucial influence on health.
Several arguments can be set forth to explain the relationship between inequality and health described
earlier. The relative deprivation suffered by people in the lowest deciles of the income distribution
may determine their exclusion from the social activities that promote or preserve health. Moreover,
as several empirical papers have pointed out, relative deprivation may be a source of persistent
stress, loss of self-esteem and chronic depression which tend to damage individuals’ health (see
Wilkinson, 2002). People compare themselves with reference groups around them (neighbours, co-
workers, friends, relatives, and so on) and may suffer from chronic stress when the comparison with
these benchmarks is unfavourable.9 This psychological mechanism adversely affects people’s health
(see, e.g. Sapolsky, 1998; Brunner and Marmot, 1999; Wilkinson, 2002). The underlying physiological
mechanism is based on the activation of hormones that affect the cardiovascular and immune systems
(Wilkinson, 2002, pp. 15–16). The mechanism through which chronic stress jeopardises the health
of individuals is very similar to economic ‘short-termism’, that is the myopic emphasis on short-
term objectives to the cost of jeopardising the achievement of longer-run objectives. In both cases,
all the available resources are mobilised to obtain a desired short-term goal even at the cost of
jeopardising the sustainability of good performance in the longer term (see Borghesi and Vercelli, 2004,
2008).
The assertion that relative income has a crucial independent impact on population health, generally
named ‘Relative Income Hypothesis’ (RIH) in the epidemiological literature, may be expressed as
follows:
Hi = ϕ2 (Y R ), ϕ2 > 0 (14)
R
where Y is the relative income that may be measured in different ways (Section 3).
The RIH is the object of a heated debate among scholars. Some authors (see, e.g. Judge, 1995;
Saunders, 1996) have argued that the RIH is inconsistent with the positive correlation between
increasing income inequality and the contemporaneous improvements in life expectancy observed
in the last decades in most industrialised countries. This criticism, however, seems to derive from
a misunderstanding of the scope of the RIH, which does not aim at explaining the long run global
trend of life expectancy (that depends on a host of factors beyond income inequality), but the negative
deviations from this positive long-run trend.
Another criticism concerns possible methodological shortcomings of the empirical studies that
support the RIH. In particular, it has been argued that most of these studies suffer from an omitted
variables problem as they do not take account of other relevant variables such as taxes, transfer
payments and household size. Some authors (e.g. Wagstaff and van Doorslaer, 2000), moreover, argue
that the results obtained from aggregate-level studies (i.e. at the population and community levels) are
not fully reliable and claim that only individual-level studies have the potential to discriminate between
the hypotheses set forth to explain the health effects of income inequality. However, more sophisticated
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 215

replications of the earlier studies adjusting for the missing variables did not change the substance of
the results obtained by earlier studies (Kawachi and Kennedy, 1997), but restricted the negative health
effects of inequality to people with medium or lower absolute income and to medium-size or larger
regions (Elstad et al., 2006).
Finally, some authors (see, e.g. Fiscella and Franks, 1997; Gravelle, 1998; Gravelle and Sutton, 2006)
have argued that only absolute income is a genuine cause of health, whereas relative income turns out
to be a spurious cause that derives from the non-linear shape of the income–health relationship. In fact,
given the concave relationship between absolute income and health (Figure 2), transferring income from
the rich to the poor (i.e. reducing inequality) improves the average health level because it increases the
health of the poor more than it reduces that of the rich.10 This third strand of criticism to the RIH is
certainly the most challenging and rigorous one. However, the observation that the relationship between
absolute income and health is non-linear implies only that inequality has an indirect effect on health
through absolute income; it does not exclude per se that inequality may also have an independent effect
on health, which is what actually emerges in several studies that control for absolute income in the
estimation model.
Therefore, although the existing results should be taken with much caution, in our opinion the
possibility that inequality has a relevant and independent health effect should not be easily dismissed.
Empirical evidence suggests that inequality engenders mistrust and hostility with negative effects
on people’s health, the more so the more incomes are perceived to be unrelated or non-proportional
to individual effort and merit. This might contribute to explain why the most egalitarian developed
countries tend to have the highest life expectancy (see Wilkinson, 2002). Uslaner (2001), in particular,
finds a high correlation coefficient (r = −0.684) between inequality and trust in a cross-country
analysis. As the author shows, this connection between the two variables holds true also in multi-
variate tests that account for economic, cultural and religious aspects that might affect the observed
levels of trust and inequality in the selected countries. As is well known, correlation per se tells nothing
about the direction of causality between the two variables. However, estimating a simultaneous equation
model to test the direction of causality between trust and inequality, Uslaner (2001) also finds that
trust has no effect on economic inequality, whereas the latter turns out to be the strongest determinant
of trust among the explanatory variables. Finally, the close relationship between income inequality
and mortality rates that is observed in cross-country studies seems to be robust to different estimation
methods, as it emerges also in time series referring to single countries, such as Russia, United Kingdom
and Taiwan.11

4.3 The Role of Other Social Factors


Empirical evidence shows that health is strongly affected by a series of social factors connected
with relative income but in part independent of it. An early example of the role of relational factors
is provided by the famous case of Roseto in Pennsylvania, a small US town built by a group of
Italian immigrants from the rural community of Roseto Valfortore in Italy (Kawachi and Kennedy,
2002). This case attracted the attention of epidemiologists since the 1950s because its inhabitants
were characterised by much better health indexes than other US towns with similar economic
and demographic characteristics. The only anomaly detected by researchers was that this group of
immigrants had managed to retain their culture of origin characterised by a much more active social
life. This anomaly gradually disappeared in the 1960s and 1970s as the culture of this town became
increasingly homogeneous with the American way of life and with it disappeared the favourable health
differentials. Social epidemiologists, alerted by the case of Roseto, argued on the basis of extensive
empirical research that there generally exists a positive and statistically significant correlation between
health indicators and social factors. We may summarise this series of social factors in terms of relational
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
216 BORGHESI AND VERCELLI

goods R

H = ϕ3 (R), ϕ3 > 0 (15)

Empirical evidence shows a clear positive correlation between interpersonal relations and health (see,
e.g. Ryff and Singer, 2000). For example, stress-related mortality of married people is significantly
lower than that of people who are widowed, divorced and single (Cornia et al., 2008). Empirical
evidence shows that, in order to withstand physiological and psychological shocks, a crucial role
is played by the intensity and quality of social relations, what is often called ‘social capital’.12 In
particular, the lack of social trust was found to be positively and significantly correlated with mortality
in the United States (Kawachi et al., 1997), with a correlation coefficient that ranges between 0.71
and 0.79 depending on the kind of social trust indicators used for the analysis. Using multi-level
regression analysis of subjective health in 40 US communities, Subramanian et al. (2002) also find
that higher levels of community social trust are associated with lower probability of reporting poor
health, and show that the health effects of social trust depend on the perception that individuals have
of it. Self-reported workplace social capital, moreover, turns out to be significantly associated with
subsequent depression and cardiovascular disease in several surveys conducted in Finland on public
sector employees who had no history of antidepressant treatment or physician-diagnosed depression at
the baseline (Kivimäki et al., 2003; Kouvonen et al., 2008). Analogously, a growing body of the medical
literature finds that hostility is positively correlated with mortality by increasing cardiovascular risk
behaviours and associated heart disease (Koskenvuo et al., 1988; Pulkki et al., 2003). In this regard,
Williams et al. (1995) estimated that mean hostility scores of 10 cities in the United States were
strongly and significantly correlated with their mortality rates after controlling for race, age, gender,
income and education level of the individuals. Similarly, using US state level data on homicide and
other categories of crime, Wilkinson et al. (2008) find that violent crime is closely related to lack of
social trust.
As for happiness, also in the case of health empirical evidence shows that other social factors such
as education and unemployment play a crucial role.
The relationship between education I and health is strongly non-linear as it increases sharply by
moving from primary to secondary education and above

H = ϕ4 (I ), ϕ4 > 0, ϕ4 < 0 (16)

In particular, most empirical studies on both developed and developing countries found that the
education level of mothers is a major determinant of the health of all the other family members,
particularly the children. Educated parents manage better the household resources in order to improve
family health, exploiting their deeper knowledge of health determinants and avoiding unhealthy
practices and lifestyle (Cornia et al., 2008, and the literature therein cited).
The empirical research agrees that unemployment U is another crucial factor of health

H = ϕ5 (U ), ϕ5 < 0 (17)

Loss of employment, especially if unanticipated and in the absence of a public safety net, and the
persistence of unemployment heavily affect health. It is also an obstacle to marriage and stable life. The
feelings of frustration and anger nurtured by unemployment tend to disrupt trust and social relations
increasing crime and violence. According to epidemiological studies, unemployed individuals face a
greater risk of mortality than the employed. Similar effects are produced by the instability of both jobs
and industrial relations (Warr, 1999).
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 217

4.4 The Influence of Environmental Externalities on Health


Besides social factors, a major and growing influence on health is exerted by the natural environment.
The World Health Organization (WHO) has estimated that bad environmental conditions are directly
responsible for about 25% of all cases of preventable illness all over the world (WHO, 1997).
Environmental risk factors, moreover, play a role in more than 80% of the diseases regularly reported
by the WHO and account for more than one-third of the disease burden among children (WHO, 2005).
We have thus to consider negative environmental externalities E as a further crucial determinant of
health13

H = ϕ6 (E), ϕ6 < 0 (18)

Atmospheric pollution is considered as the main cause of the large increase of respiratory diseases
observed in recent years. Some particularly volatile pollutants such as fine dust (PM10 ), nitric oxide
(NOx ) and sulphur dioxide (SO2 ) – discharged by cars traffic, heating and manufacturing – can penetrate
into the bronchioles, provoking asthma, bronchitis and emphysema (Kunzli et al., 2000; Pope et al.,
2002).14 Besides respiratory conditions, atmospheric pollutants are often responsible for cardiovascular
diseases because, once inhaled, they are carried around the body by blood.
As for water pollution, the concentration of faecal coliform bacteria in water lacking efficient
treatment is an index of pathogenic agents responsible for diarrhoea, cholera, hepatitis, typhoid fever
and other illnesses of the digestive system. Several studies (e.g. WHO, 1997) have estimated that 90%
of all cases of these diseases can be ascribed to the lack of clean water and to inadequate sanitation,
a percentage that rises up to 94% in the case of diarrhoeal diseases. The latter are estimated to cause
about 1.7 million deaths annually (WHO, 2005), particularly among children in developing countries
(where 95% of water is untreated).15
Another factor of water pollution that has serious consequences for human health is the presence in
water of heavy metals (such as lead, cadmium, mercury, arsenic and nickel) and polluting chemical
products (such as PCB, DDT and dioxins). People ingest these elements by drinking water since
they are difficult to remove under normal treatment processes, or when they eat fish where metals
can accumulate. Various studies demonstrated that some heavy metals, such as nickel, cause serious
damage to the nervous system; others, such as lead, mercury and arsenic, harm liver and kidneys.16 All
heavy metals and many chemical pollutants are also thought to be responsible for tumour formation.
Furthermore, water pollution in combination with atmospheric pollution can modify the habitat of some
ecosystems (temperature, humidity, vegetation density, and so on), and this can enhance the survival
and spreading of insects that are particularly harmful because of the diseases they may carry. This is the
case of mosquitoes, which transmit various diseases including malaria. This serious disease is thought
to be responsible for over half a million deaths every year (particularly among children aged under 5
years) and is becoming an increasingly serious problem, especially in sub-Saharan Africa where 90%
of the world’s malaria cases are concentrated (WHO, 1997 and 2005). It is estimated that a large share
of malaria (about 42% of all cases according to WHO, 2005) can be ascribed to readily modifiable
environmental factors, such as land use, irrigation and agricultural practices, therefore much of the
burden of the disease could be easily prevented by improving the environmental management of the
areas involved.
Finally, many chemical, biological and radioactive pollutants tend to settle on the soil, contaminating
both the crops planted there and the resultant agricultural products. In addition, soil pollution damages
the health not only of farmers who work contaminated land and of children playing there, but also of
the surrounding population since dust from the polluted area can be carried elsewhere by the wind.
Direct contact with contaminated soil and with the numerous microbes and parasites contained in it is
particularly harmful for children who are extremely vulnerable.17
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
218 BORGHESI AND VERCELLI

4.5 Other Factors: Medical Technology and Genetic Factors


Medical technology T played a crucial role in the progressive improvement of the health indexes over
the last century
H = ϕ7 (T ), ϕ7 > 0 (19)
Thus, for instance, the higher life expectancy in old age that has been observed in OECD countries
is largely due to improved access to quality health services and medical progress, particularly for
cardiovascular diseases (OECD, 2005). Conversely, the dramatic increase in Southern Africa of the
diffusion of AIDS (and the consequent decrease of life expectancy) since the late 1980s is largely
tied to the lack of diffusion also in these countries of specific medicines (antiretroviral therapy)
aimed at preventing the evolution of the disease. In Botswana, for instance, life expectancy has
decreased from 65 years in 1985–1990 to 37 years in 2000–2005, whereas in Southern Africa
as a whole (that encompasses most of the countries highly affected by the HIV/AIDS epidemic)
life expectancy has fallen from 61 to 48 years over the same period (United Nations, 2005). The
HIV/AIDS epidemic is still spreading and its impact on life expectancy in these countries is expected
to keep on worsening for the next few years before lessening. The United Nations (2005) estimate that
the turning point in life expectancy should hopefully occur when the ongoing antiretroviral therapy
will reach an ever increasing share of the persons who need it, resulting in higher survivorship
for people with HIV and lower infectious rates in the countries most heavily affected by the
disease.
Also genetic factors G may have a sizable impact on health
H = ϕ8 (G) (20)
We have to distinguish two main factors: mutation of genes and polygenic inheritance, that is the
specific combination of normal genes that confers a bias towards specific chronic diseases, such as
high blood pressure, diabetes and cancer. Research has found about 4,000 mutant genes that may
cause diseases such as sickle-cell anaemia, cystic fibrosis and Huntington’s disease (Tarlov and St.
Peter, 2000, p.x). Epidemiologists maintain that the relative incidence of these diseases is fairly low
as it did not exceed a value around 5% of the total (Tarlov and St. Peter, 2000). In the case of
polygenic inheritance, for a chronic disease actually to manifest itself, concomitant circumstances have
to concur, such as health-damaging behaviours or socio-economic factors. We may thus consider its
impact important but only in the distribution of chronic diseases rather then in their aggregate incidence
on population health.

4.6 Concluding Remarks


The main determinants of health considered in the epidemiological literature do not exclude each other.
We can thus summarise the function that explains the behaviour of the health status in the following
way:
H ∗ = (Y , Y R , R, U , I , G, E, T , W ) (21)
where we assume that the partial derivatives for each factor have the same signs discussed earlier. The
capital letters that appear in the argument of the function  should be interpreted as in the happiness
case as vectors of variables some of which have a flow dimension and others a stock dimension. So
the health of people depends on their absolute income but also on their wealth, on relative income
and wealth, on alterations to genetic traits (because of cosmic or anthropogenic radiations) and the
genetic patrimony of individuals, on flow environmental externalities (e.g. polluting emissions) and
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 219

accumulated stocks of their effects (such as the concentration of greenhouse gases in the atmosphere),
on relational goods and social capital, on training and degree of education, change of working status
and curriculum.
Also in this case, of course, other health determinants that are significant for particular countries and
periods may be included among the arguments of the function. However, we preferred to include only
the factors that have been systematically analysed in the epidemiological literature and have found a
fairly robust empirical support.
A particularly important factor excluded from our analysis is personal lifestyle. As is well-known,
individual health is heavily influenced by personal lifestyle and behavioural habits such as the time
devoted to sport, open-air activity and periodical medical checks, nutritional habits, smoking, and
so on. These aspects are only partially related to per capita income. Therefore, the changes in life
expectancy that have been observed in the last decades at the world level cannot be fully ascribed
to variations in per capita income, but are also likely to mirror significant variations in lifestyle and
behavioural habits. In this regard, it should be noted that an increase in per capita income is sometimes
related to a rise in health-damaging behaviours that may partially counterbalance the positive effects
of higher income. Poor diet and excessive tobacco and alcohol consumption, for instance, may explain
why in some Central and Eastern European countries (e.g. Hungary and Slovakia) male life expectancy
increased only slightly or remained unchanged in the last few decades despite an increase in per
capita income (OECD, 2001). In other eastern European countries, life expectancy actually declined
and remained significantly lower than the levels recorded before the transition to a market economy,
despite a resurge in per capita income in the last decade (OECD, 2005; United Nations, 2005). In
the Russian Federation, for instance, average life expectancy fell from 69.17 to 65.55 years in the
period 1989–2006, while in Ukraine it decreased from 70.5 to 68 years over the same period (World
Bank, 2008). In these two countries life expectancy kept falling or remained unchanged even during
the last decade in which per capita GDP started increasing again after reaching its lowest level in
1998.18 Moreover, some developed countries show lower average health status than some developing
countries due to sedentary life habits and (over)consumption of unhealthy fast-food as a consequence
of stressful work timetables.19 Finally, an increase in the number of anorexia cases has been observed
in most developed countries, especially in the young generation that is particularly exposed and
sensitive to the beauty standards proposed by the mass-media that induce severely health-damaging
behaviours. This represents a particularly interesting new example of a complex link between health
and happiness that would deserve further analysis in the future. However, we do not include this set
of factors in the health function since for the time being it would be difficult to measure their impact
with a reliable index. As for their impact on happiness we can assume that it is exerted through
health.

5. The Link Between Happiness and Health


As pointed out earlier, most of the empirical regularities that emerge from the literature surveyed
before are still in search of a fully convincing theory, as it is typical of the empirical literature in the
early stages of development. Many of these empirical regularities are still basically mere statistical
correlations and we do not really know to what extent they are spurious or are susceptible of sound
causal interpretation. To this end, a wealth of theoretical hypotheses have been put forward on which are
the crucial independent variables but causal inference is still hindered by unsolved problems deriving
from complex interaction of endogenous variables and multi-collinearity.
Although causal analysis is in its infancy, it is important to draw attention on the common features
of the two parallel literature streams surveyed earlier.
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
220 BORGHESI AND VERCELLI

5.1 Twin Paradoxes


The strong analogy between the two research strands emerges immediately from a comparison between
the comprehensive functions (12) for happiness and (21) for health. It is striking that two streams of
literature pursued independently by different groups of researchers focused mainly on the same kernel of
systematic factors finding similar patterns of correlation for each of them. Of course, the emphasis, the
language and the conceptual framework are often different but it is difficult to find serious divergences
in the basic results of the analysis. The only substantial differences emerging from a prima facie
comparison of the two general equations are the absence of aspirations and the presence of the health
technology in the health equation. As for frustrated aspirations, the epidemiological literature confirms
its crucial role in the psycho-physical health of individuals (Marmot, 2004) but its role is generally
captured by happiness and by some of the other variables that appear in the health function. As for the
role of the health technology, its effects on happiness are captured by the presence of health among the
arguments of the happiness equation and does not need to appear explicitly. Analogously, the effects
of technical change in general (i.e. not circumscribed to medical technical change) are largely captured
by income in the happiness equation so that technology does not need to be considered explicitly in
both the health and happiness equations.
By comparing the determinants of happiness and health, we may shed further light on their mutual
relationship as well as on the twin paradoxes, namely, on the finding that neither health nor happiness
increases much after medium-income levels. The cross-section correlation between per capita income
and health is very similar to that between per capita income and happiness. In both cases, we have
a concave correlation that is quite steep for low cohorts of income, becomes progressively less steep
for higher incomes and fades away after a certain threshold. The thresholds are slightly different but
are in both cases surprisingly low: between $4,000 and $5,000 in the case of health and between
$10,000 and $15,000 in the case of happiness (see retro Sections 3 and 4). The threshold for health is,
therefore, fairly lower than that for happiness. In addition, in the case of health the correlation looks
steeper, almost vertical below the threshold, and almost flat beyond it, while in the case of happiness
its marginal rate of change is more gradual as income grows. This empirical evidence is consistent
with the hypothesis that the hierarchy of needs plays a crucial role in determining both happiness
and health. Up to the threshold mentioned above, an increase of income removes the constraints to
the satisfaction of a few basic needs (food, dwelling, access to safe water and sanitation) that have
a strong impact on both health and happiness. This threshold is quite low because the satisfaction
of these basic needs is strongly subsidised in most countries. Once these needs are satisfied, new
needs emerge whose satisfaction improves health only marginally, but may still significantly improve
happiness up to its own threshold. Beyond the latter, a further increase in absolute income does
not seem to have a significant impact on either happiness or health (Frey and Stutzer, 2002). In
both cases, after the specific threshold, the crucial causal role is taken over by relative income and
social factors. As soon as the problem of survival is solved, the attention focuses on the perceived
social status. This does not imply that social status is not important in poor communities such as the
favelas of Brasil or the remote villages of Equatorial Africa or rural China (see Clark et al., 2006).
However, up to the health threshold, the effects of relative deprivation are overwhelmed by the effects
of absolute deprivation. After this threshold, relative deprivation becomes the crucial factor. The basic
aspirations are set by the will to ‘keep up with the Joneses’ while the Joneses try hard to keep the
distances. Relative deprivation and frustrated aspirations generate stress and are sometimes sources
of chronic and psychosomatic maladies, as well as of persisting feelings of unhappiness. Another
important source of happiness and health are the social relations entertained within the family, with
friends and other members of the communities to which the individual belongs by choice or birth.
The deterioration of relational goods is thus another important source of unhappiness and health
degradation.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 221

5.2 Happiness, Health and the Environment: The Twin Paradoxes and the Environmental
Kuznets Curve
As previously pointed out, environmental degradation too plays a crucial role for the explanation of the
twin paradoxes. The environmental negative externalities of growth force defensive expenditures that
are to be subtracted from happiness-improving consumption.20 Moreover, they adversely affect both
happiness and health in several ways as described earlier, so that one can reasonably expect a negative
correlation between environmental degradation and health (happiness). One may wonder whether this
is consistent with the literature on the so-called environmental Kuznets curve (EKC). As is well known,
following the results originally appeared in a report of the World Bank (1992), in the early 1990s several
studies (e.g. Shafik, 1994; Selden and Song, 1994; Grossman and Krueger, 1995) have observed an
inverted-U shaped relationship between environmental degradation and per capita income. This suggests
that environmental degradation first increases at low-income levels, reaches its peak at middle-income
levels and then decreases at high-income levels as per capita income overcomes a given threshold.
The bell-shaped path of the EKC seems prima facie inconsistent with the relationship between health
(happiness) and per capita income described above, according to which health (happiness) increases
for low-income countries and reaches its maximum for middle-income countries. In other words, if an
EKC does exist then both environmental degradation and health (happiness) will tend to increase as per
capita GDP grows shifting from poor- to middle-income countries, so that one should observe a positive
correlation between environmental degradation and average health (happiness), rather than a negative
one as argued earlier. If an EKC is empirically observed, moreover, environmental degradation is much
higher in middle-income countries than in high-income countries, although they both experience similar
levels of happiness and health (Figures 1 and 2), which seems counter-intuitive a priori. However, the
evidence supporting the EKC is at best very weak. In the absence of a single criterion of environmental
quality, many environmental indicators have been used in the literature to test the EKC hypothesis.
Among them, only a small subset tend to follow an inverted-U relationship with per capita GDP. In
particular, some local air quality indicators (e.g. sulphur dioxide, suspended particulate matters, carbon
monoxide and nitrous oxides) show a strong (but not overwhelming) evidence of an EKC. However,
when emissions of air pollutants have little direct impact on the population, the literature generally
finds no evidence of an EKC. In particular, both early and recent studies find that emissions of global
pollutants (such as carbon dioxide (CO2 )) either monotonically increase with income or start declining
at income levels well beyond the observed range (cf. Borghesi, 2001, for a critical review of the
literature and Dinda, 2004 for an updating of the evidence at disposal). In addition, when a more
comprehensive measure of environmental degradation such as the ecological footprint is used to test
the EKC hypothesis, the literature finds no evidence of an EKC (Caviglia-Harris et al., 2009).
Even when an EKC is empirically observed, there is still no agreement in the literature on the
income level at which environmental degradation starts decreasing.21 Some contributions, moreover,
have questioned the existence of the EKC even for those indicators that seem to follow this pattern
and argued that the EKC might be a ‘statistical artefact’ (Vincent, 1997) of the econometric approach
adopted so far. Given the scarcity of long time-series of environmental data, in fact, most studies
have used a cross-country approach. However, this approach might be misleading since it may merely
reflect the juxtaposition of two opposite trends of environmental degradation (increasing in developing
countries and decreasing in industrialised ones), rather than describe the evolution followed by a
single economy over time. In fact, single-country studies that have examined the environment–income
relationship over time generally find no evidence of an EKC (Vincent, 1997; de Bruyn et al., 1998; Roca
et al., 2001; Deacon and Norman, 2006). Finally, as several authors (e.g. Cole et al., 1997) have pointed
out, both cross- and single-country studies are based on reduced form models, therefore they reflect
correlation rather than a causal mechanism and can give no indication about the direction of causality,
namely whether growth affects the environment or the other way around. In reality, environmental
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
222 BORGHESI AND VERCELLI

quality may have several feedback effects on income growth (Stern et al., 1996; Shen, 2006) because
it can affect the productivity of both workers and natural resources. Hence, a simultaneous-equations
model would certainly be more appropriate to deepen our understanding of the environment–income
relationship and its implications for health and happiness.
These results, together with the existence of data quality and comparability problems as well as
limitations in econometric techniques, cast serious doubts on the evidence in favour of the EKC and
on the reliability of the indications emerging from cross-country studies in this context. In our opinion,
therefore, policy makers should not take the alleged shape of the EKC to conclude that growth will
automatically solve the ecological problems it causes in the early stages of development. On the
contrary, the rising environmental problems that are observed at both the local and the global level
might provide a reasonable foundation underlying the twin paradoxes in terms of health and happiness.

5.3 Happiness and Health: The Role of Ageing


Since long, the epidemiological literature has found a strong correlation between measures of self-
reported happiness and several health indexes, such as length of life (Palmore, 1969), heart disease
(Sales and House, 1971), suicide (Koivumaa-Honkanen et al., 2001) and strokes (Huppert, 2006).
However, as soon as we compare the post-war trends of the most comprehensive health indexes (life
expectations and mortality rates) with the trends of reported happiness in industrialised countries, the
two trends turn out to be weakly related. As is well known, life expectations and mortality rates
improved continuously in most countries after WWII, similarly to per capita income, while reported
happiness did not follow a similar path.22 Also in this case the weak, sometimes slightly negative,
correlation between health and reported happiness apparently runs against the expectations and the
opinion repeated by most people that health is a major determinant of happiness.
One may wonder why the continuous improvement of health after WWII in developed countries did
not translate in increasing self-reported happiness. A possible explanation is that the progressive increase
in life expectancy and reduction of mortality rates determined a continuous ageing of the population
that is negatively correlated with subjective happiness. For instance, if higher life expectancy is mainly
due to people living longer in poor health conditions kept alive by specific medicines, then stagnant
happiness could be easily explained by lower average health that counterbalances all the other positive
driving forces. In fact, the importance of health is declared to increase with age while self-reported
health declines, which should reflect negatively on happiness. We believe that this causal link may have
had some influence on stagnating happiness in the post-war period but its impact, evaluated in the light
of the existing empirical evidence, is unlikely to have been decisive for at least two reasons. In the
first place, longer life expectancy does not seem to be correlated with poorer health conditions among
the elderly. On the contrary, most measures of average health tended to improve over the last decades
among the older sections of the population. For instance, using repeated longitudinal surveys of the
older population over the period 1984–2000, Crimmins et al. (2009) find that in the US life expectancy
free of disability increased over a 10-year period by 0.6 of a year in the later cohort (the population 70
years of age and older).23 A similar decline in the prevalence of disability among elderly persons over
the last two decades emerges also from other recent studies (Waidmann and Liu, 2000; Freedman et al.,
2004; Manton et al., 2006). More precisely, empirical evidence shows that less severe disability started
declining in the 1980s, while more severe disability began to fall from the 1990s onwards (Crimmins,
2004; Freedman et al., 2004). The amount of time the elderly can expect to live without disability
has been increasing not only in the United States, but also in many other industrialised countries (e.g.
France, Belgium, Taiwan, Italy, the Netherlands and Switzerland).24
In the second place, existing empirical evidence suggests that in most developed countries happiness
declines from youth to working age but increases again, although moderately, since around the age of
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 223

retirement (Frey and Stutzer, 2002). Clark and Oswald (1994), for instance, have found a U-shaped
relationship between happiness and age in Great Britain with happiness being lowest in a person’s mid-
thirties. Using a three-point categorical measure of overall happiness to avoid the limitations of the
disutility scale adopted by Clark and Oswald (1994), Gerdtham and Johannesson (2001) have observed
a similar relationship for Sweden, although the minimum happiness turns out to be slightly postponed
(being lowest for the age-group 45–64 years). Similar results apply to Australia, New Zealand and
the United States (Deaton, 2008), with the latter country showing the same score for life satisfaction
at 20 as at the age of 60.25 The observed U-shaped relationship between happiness and age may be
related to increasing free time of senior people, decreasing responsibilities, downsizing of aspirations
and, maybe, natural selection of the healthier and happier individuals. We have to conclude that the
negative effect exerted by ageing on health is likely to be more than compensated by other factors
correlated with age.
As argued earlier, the ageing of the population is not sufficient per se to explain the observed
relationship between happiness and health, therefore we have to look for alternative explanations.
Among them, a crucial role is certainly played by the shortcomings of the objective health indexes at
disposal. There are compelling reasons to suggest that the comprehensive health indexes mentioned
earlier (life expectations and mortality rates) do not reflect well the effects exerted by health on
subjective happiness. The latter, in fact, seems to depend much more on the subjective perception that
individuals have of their health than on their effective health status.
The correlation between subjective indexes of health and happiness is generally quite high. For
example, performing a cross-country regression of average life satisfaction on a set of explanatory
variables including health satisfaction, Deaton (2008) finds that the latter variable has a large and
statistically significant coefficient. Using a cross-sectional survey of community-dwelling older adults,
moreover, Angner et al. (2009) find that subjective health measures are generally better predictors
of happiness than objective health measures. This confirms the pre-eminent subjective importance
attributed by individuals to health in their self-assessment of happiness. Kahneman and Riis (2005)
also find a high correlation coefficient (r = 0.85) between subjective indexes of health and happiness
in the 18 OECD countries taken into account. On the contrary, subjective health turns out to be
only weakly correlated with the objective measures of health (such as life expectancy and mortality
rates) in the same OECD countries (Kahneman and Riis, 2005). This interesting result has been
further confirmed by Deaton (2008). Using the Gallup World Poll dataset (based on the identical
questionnaire conducted by the Gallup Organization in 132 countries in 2006), the author finds that
health satisfaction is uncorrelated to objective health measures such as life expectancy, its increase over
the period 1990–2005 or even the prevalence of HIV infection. In Deaton’s analysis, this lack of a link
between objective and subjective measures of health turns out to be robust also to different estimation
models and more sophisticated econometric techniques.26 These results, moreover, are consistent with
previous findings by Sen (2002) and Chen and Murray (1992) who observed that self-reported health
measures are often better in countries with poorer health conditions, probably because people tend to
adapt and get used to hardships, modifying their perception of what is good or bad health.27
Summing up, a plausible explanation of the observed decoupling between life expectancy and self-
reported happiness in the post-war period may rely on the fact that subjective health and subjective
happiness depend not only on the length of life but crucially also on its quality. A long life is not
necessarily a happy life.28

6. The Twin Paradoxes and their Policy Implications


The twin paradoxes and the underlying driving forces discussed earlier suggest that economic growth
is no longer the major determinant of happiness and health beyond a certain per capita income level.
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
224 BORGHESI AND VERCELLI

This observation should not be taken to imply that income growth per se is irrelevant or, even
worse, harmful for happiness and health in industrialised countries. However, when income is high
enough, what matters for these two variables is not so much the rate of growth but rather the kind of
growth pursued by the country. This calls, therefore, for social and environmental policies that aim at
promoting happiness and health not only through income growth but also through the additional factors
that jointly affect these two variables.
Among these policies, a crucial role is played – in our opinion – by those who support investments
in research and development (R&D) and technological progress. As is well known, in fact, such
investments tend to promote economic growth (Romer, 1990; Aghion and Howitt, 1992; Griffith,
2000), which may bring about a rise of happiness and health in the developing countries, that lie
on the increasing part of the curves shown in Figures 1 and 2. Specific R&D investments in the
health and environmental sectors, moreover, may increase happiness and health also in the developed
countries, that lie on the horizontal section of the curves where economic growth stops playing a major
role for these two variables. As a matter of fact, higher investments in low-impact, environmental
friendly technologies tend to improve the environmental quality of the country, which is likely to
increase in its turn the happiness of the citizens living there (for the better environmental conditions
that they can enjoy) as well as their physical health (for the consequent reduction of the pollution-
related diseases). Similarly, devoting higher R&D resources to the health sector for the development
of new technologies and better infrastructures can significantly improve the average health status of
the population, particularly if this is accompanied by appropriate policies that promote the effective
access of the poorest sectors of the population to the most recently developed techniques. A better
access to modern health technologies and infrastructures, moreover, may provide individuals with the
feeling of having better chances to be properly cured (and cared for), which is likely to enhance also
their self-reported happiness, as suggested by the empirical evidence reported earlier.
Also redistribution policies that transfer financial resources from the rich to the poor sectors of the
population can turn out to be very important for both health and happiness. The concave relationship
between health (happiness) and per capita income implies that a lower income inequality is associated
with higher levels of health (happiness), the more so the more polarised is the income distribution. By
reducing income inequality, therefore, redistribution policies can increase both happiness and health.
In addition, although it is still quite controversial, income inequality has been found in many studies
to affect negatively also subjective well-being and health of lower-rank people and to contribute to the
depletion and deterioration of social capital in a population. The progressive redistribution policies,
therefore, may improve happiness and health also by reducing the feelings of frustration (on the part
of the relatively poor), social insecurity and related mental stress (on the part of the relatively rich)
and life quality dissatisfaction that are often reported in countries that experience high inequality rates
(e.g. Alesina et al., 2004; Graham and Felton, 2005; Biancotti and D’Alessio, 2007).
In this regard, an important contribution to jointly improve happiness and health may also come
from the policies that tend to protect and promote social capital. In particular, any policy that favours
social inclusion, transparent institutions and social networks can counterbalance the decline of social
capital that may derive from several features of modern life style, such as worsening urban life quality
conditions, stressful work and increasing time constraints due to the expansion of the economic activity
and rising diffusion of home entertainments that tend to replace relational goods (Antoci et al., 2007,
2009). The same applies to the policies that aim at supporting the family that can significantly increase
the health and happiness of the family members.
The observations set forth earlier can have important policy implications also for the heated debate on
the role of economic growth for the individual and social well-being. In general, the twin paradoxes and
the possible underlying mechanisms discussed above seem to suggest that we should try to emancipate
economic policy from the unjustified obsession for income growth since the quality of income turns
out to be much more important than its crude quantitative measures adopted in national accounts.
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 225

7. Concluding Remarks
In this paper, we aimed at contributing to a strategy of cross-fertilisation between two rapidly growing
research fields dealing respectively with the socio-economic determinants of happiness and health.
Although these two research streams developed separately, they are characterised by strong structural
analogies that are rich of potential insights and call for more interaction between them. The empirical
literature confirms that health is a major determinant of subjective happiness and that the converse is
also true. We have then ascertained that the other main factors affecting both happiness and health have
a common kernel and the specification of their causal influences is very similar even in their functional
form, although much caution is needed to infer the direction of causality due to possible endogenous
variables and multi-collinearity problems. In particular, as it emerges from the literature reviewed in
this work, the progressive erosion of social capital and environmental goods can also contribute to
explain the mechanism underlying the twin paradoxes.
Despite this strong analogy between the behaviour of happiness and health, the two variables
turn out to be weakly correlated in developed countries. In these countries most health indexes (in
particular, life expectancy and mortality rates) progressively improved in the post-war period while the
subjective indexes of happiness did not improve in the same period. A possible explanation of this
prima facie surprising result relies on the shortcomings of the general indexes of well-being (income)
and health (life expectancy and mortality rates) since both classes of measure do not take into account
the quality of life that plays a crucial role for the self-reported happiness of individuals and their
health.
The cross-fertilisation between the rapidly growing literatures on the socio-economic determinants
of happiness and health may shed new light also on the complex issues raised by the choice of a
policy strategy meant to improve happiness and health. The assumption of objective policy targets
different from their subjective counterpart requires a sound understanding of the reasons for such
normative deviations and a wide political consensus based on such understanding. Although these
problems are far from being solved, we cannot ignore the policy implications of the happiness
literature.
A relevant inference that can be drawn from the literature surveyed in this paper is that economic
growth per se is loosing importance for the well-being of citizens in developed countries and should
recede from its role of primary policy goal. This goes against the mainstream conviction that economic
policy should continue to focus on the maximisation of income growth. The twin paradoxes, however,
suggest that this policy target is misleading because the GDP statistics do not take account of crucial
factors of well-being. The real paradox is actually that the well-known shortcomings of GDP statistics
are still systematically ignored by policy authorities, mass media, many scholars and large parts of
public opinion.
As for income distribution, the happiness literature tends to reject the hypothesis that inequality may
represent an incentive for growth or an unpleasant but inevitable collateral consequence of growth.
Redistributing income from the rich to the poor would reduce both income and health inequalities
improving the average health of the population, since it would benefit the health of the poor much
more than it would affect that of affluent people. A policy intervention of this kind would require
a continuous and consistent policy strategy directed to improve the distribution of income, such as
progressive taxation, social transfers to disadvantaged people and facilitated access to services to
poorer people.
Finally, the analysis of the twin paradoxes suggests that policies directed to invest in environmental
and social capital, as well as in education, culture and creative goods are likely to improve both
health and happiness of the population. For this purpose, in our opinion it is particularly important to
invest increasingly more in the future in the necessary material and immaterial infrastructures aimed
at enhancing social capital. As for material infrastructures, an urban design characterised by many

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
226 BORGHESI AND VERCELLI

squares and meeting places, an intense cultural life, and a salubrious environment would improve
both happiness and health. As for immaterial infrastructures, the availability of a network of services
and social facilities may provide an important contribution to reduce stress and morbidity across the
population.

Acknowledgements
The authors wish to thank two anonymous referees for their useful comments and suggestions on an earlier
draft of this paper. All the remaining shortcomings are our responsibility.

Notes
1. We thank an anonymous referee for drawing our attention on this aspect. To avoid possible confusions,
in what follows we will always spell out the health indexes we are referring to (mainly life expectancy
and mortality rates), distinguishing between ‘objective’ and ‘subjective’ only when specifically needed
to clarify the argument.
2. See the survey by Hansen and Østerdal (2006) for a thorough discussion on the controversial
methodological issues concerning the use of quality-adjusted life years models.
3. The regression line in the diagram describes how a logarithmic curve fits the data.
4. According to the traditional point of view of economics, the total utility of an individual depends mainly
on consumption and therefore is a growing function of per capita income. The slope of the function is
progressively decreasing according to the law of diminishing marginal utility that is generally explained
in terms of satiation and hierarchy of needs.
5. The set of personal relations that support relational goods define the civic society (or community) not
to be confused with society in its strict sense or the set of impersonal relations mediated by the market
and/or the political process. Economics has traditionally focused on society fully neglecting community,
so loosing the opportunity of understanding the strong influence that relational goods have on the
working of the economy and on the happiness of citizens (see Bruni, 2006).
6. According to Antoci and Bartolini (2004) these negative externalities are the mainspring of economic
growth.
7. Kaplan et al. (1996) found that the correlation coefficient between the age-adjusted mortality rates
and the income proportion that goes to the least well off 50% of the population is high and basically
unchanged when median income is also taken into account among the explanatory variables, shifting
from 0.62 to 0.59 with p < 0.001 in both cases. On the contrary, the correlation coefficient between
total mortality and median income is much lower and falls drastically from 0.28 (p < 0.05) to 0.06 (p
> 0.05) when adjusted for income inequality.
8. Lynch et al. (2000), for instance, observed that higher inequality has been related to lower mortality
rates in Britain during the period 1962–1990.
9. Deaton (2002) argues that this psychological mechanism plays a crucial role in causing stress to the
agents and sets up a model assuming that each individual’s stress is proportional to the total amount of
income that goes to richer people in the community.
10. The inference of mistaken assertions about individual-level relationships from aggregate data is often
called ‘ecological fallacy’ in this literature.
11. Much of the relevant research has been collected in one volume (Kawachi et al., 1999).
12. Social relations are generally conceived as flow variables. However, they leave a persistent trace in terms
of dispositions of people to new encounters and of facilities that reduce their costs. These persistent
effects may accumulate in a stock that may be interpreted as a crucial component of social capital. See,
for example Pugno (2007).
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 227

13. In this section, we focus on the direct health effects of environmental degradation. See Section 5 for a
further discussion on the indirect link between health and the environment that takes economic growth
and the so-called environmental Kuznets curve into account.
14. WHO (1997) estimates that atmospheric pollution is also directly responsible for 2% of cases of cancer.
15. It has been estimated that 88% of deaths due to intestinal diseases involve children under 15 years of
age, a much higher incidence than the average number of deaths under 15 years of age due to other
diseases (WHO, 1997).
16. See, for example Conservation Foundation (1992).
17. This may contribute to explain, for example, the high incidence of neonatal tetanus in the poorest areas
of the developing countries.
18. In the period 1998–2006, life expectancy continued to fall in the Russian Federation (from 66.78 to
65.55 years), whereas it remained basically unchanged in Ukraine (around 68 years) despite a remarkable
increase of their per capita GDP (by 73.42% and 75.52% over the observed period, respectively) (cf.
World Bank, 2008). Notice, however, that this negative performance of life expectancy in the Russian
Federation and Ukraine is largely due also to the increasing environmental degradation observed in the
two countries.
19. As is well known, at the world level the number of individuals affected by obesity has recently overcome
those suffering malnutrition.
20. Several works (e.g. Bartolini and Bonatti, 2002, 2003; Antoci and Bartolini, 2004; Antoci et al.,
2008) show that the environmental defensive expenditures that the agents perform to self-protect from
environmental degradation may give rise to negative welfare effects for the society as a whole. See
Antoci and Borghesi (2010) for an extensive review of the literature on this issue and an application to
the North–South context.
21. See Borghesi (2001) for a comparison of the different turning points that emerge in different studies on
single environmental indicators.
22. On the contrary, some specific health indexes that are strictly correlated with self-reported unhappiness,
such as frequency of depression and suicides, increased progressively in the post-war period (Kawachi
and Kennedy, 2002).
23. The existing literature on disability trends in the older population (cf. Pope and Tarlov, 1991; Waidmann
and Liu, 2000) defines as ADL disabled a person who has difficulty and requires help with any of
the following basic activities of daily living (ADL): walking across a room, eating, dressing, bathing,
transfer (getting in/out of chairs), and toileting. Similarly, it defines as IADL disabled a person who
encounters difficulties to perform the following instrumental activities of daily living (IADL): using
the telephone, doing light housework, doing heavy housework, preparing meals, shopping for personal
items, and managing money.
24. See Waidmann and Manton (1998) for a comprehensive review of the evidence on this issue, both for
the United States and internationally.
25. Results on the happiness–age relationship differ, however, according to the country taken into account.
Life satisfaction, for instance, tends to decline with age in other countries, particularly in Eastern Europe
and the former Soviet Union. See Deaton (2008) for a country-specific explanation of the trend observed
in these countries.
26. Health satisfaction, on the contrary, turns out to be highly related to what people think of their health
care system (Deaton, 2008). Since the degree of confidence in the health care system is itself subjective,
this further confirms the high correlation between subjective evaluations in different domains (i.e. life
satisfaction, health satisfaction and health care system satisfaction).
27. See Layard (2006) and Deaton (2008) for the ongoing debate on the pros and cons of life
and health satisfaction measures and their capacity of adequately reflect objective conditions of
health.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
228 BORGHESI AND VERCELLI

28. This is well known since long. In the island of Luggnagg visited by Gulliver in one of his travels, some
of the inhabitants (Struldbrugs) are very unhappy because they become older and older but cannot die
(Swift, 1726).

References
Abdallah, S., Thompson, S., Michaelson, J., Marks, N. and Steuer, N. (2009) The (un)Happy Planet Index
2.0. Why good lives don’t have to cost the Earth. London: New Economics Foundation.
Aghion, P. and Howitt, P. (1992) A model of growth through creative destruction. Econometrica 60: 323–351.
Alesina, A., Di Tella, R. and McCulloch, R. (2004) Inequality and happiness: are Europeans and Americans
different? Journal of Public Economics 88: 2009–2042.
Angner, E., Ray, M., Saag, K. and Allison, J. (2009) Health and happiness among older adults. A community-
based study. Journal of Health Psychology 14: 503–512.
Antoci, A. and Bartolini, S. (2004) Negative externalities and labor input in an evolutionary game.
Environment and Development Economics 9: 591–612.
Antoci, A. and Borghesi, S. (2010) Environmental degradation, self-protection choices and coordination
failures in a North-South evolutionary model. Journal of Economic Interaction and Coordination 5:
89–107.
Antoci, A., Sacco, P.L. and Vanin P. (2007) Social capital accumulation and the evolution of social
participation. Journal of Socio-Economics 35: 128–143.
Antoci, A., Borghesi, S. and Galeotti, M. (2008) Should we replace the environment? Limits of economic
growth in the presence of self-protective choices. International Journal of Social Economics 35: 283–297.
Antoci, A., Sabatini, F. and Sodini, M. (2009) The fragility of social capital. FEEM Working Paper 2009.16,
Milan: Fondazione Eni Enrico Mattei.
Argyle, M. (1996) The Social Psychology of Leisure. New York: Penguin.
Bartolini, S. and Bonatti, L. (2002) Environmental and social degradation as the engine of economic growth.
Ecological Economics 41: 1–16.
Bartolini, S. and Bonatti, L. (2003) Endogenous growth and negative externalities. Journal of Economics 79:
123–144.
Biancotti, C. and D’Alessio, G. (2007) Inequality and happiness. ECINEC – Society for the Study of
Economic Inequality, Working Paper No. 2007-75.
Blanchflower, D.G. and Oswald, A.J. (2004a) Money, sex and happiness: an empirical study. Scandinavian
Journal of Economics 106: 393–415.
Blanchflower, D.G. and Oswald, A.J. (2004b) Well-being over time in Britain and the USA. Journal of
Public Economics 88: 1359–1386.
Borghesi, S. (2001) The environmental Kuznets curve: a critical survey. In M. Franzini and A. Nicita (eds),
Economic Institutions and Environmental Policy (pp. 201–224). Aldershot: Ashgate. Previously published
as Nota di Lavoro No. 85.99, Fondazione ENI Enrico Mattei, Milan.
Borghesi, S. and Vercelli, A. (2004) Globalisation, inequality and health. International Journal of Global
Environmental Issues 4: 89–108.
Borghesi, S. and Vercelli, A. (2008) Global Sustainability. New York: Palgrave-MacMillan.
Bottan, N.L. and Perez Truglia, R.N. (2008) Deconstructing the Hedonic Treadmill. MPRA Paper 10268,
Munich: University Library of Munich.
Brickman, P. and Campbell, D.T. (1971) Hedonic relativism and planning the good society. In M.H. Apley
(ed.), Adaptation-Level Theory: A Symposium (pp. 287–302). New York: Academic Press.
Brickman, P., Coates, D. and Janoff-Bulman, R. (1978) Lottery winners and accidents victims: is happiness
relative? Journal of Personality and Social Psychology 36: 917–927.
Bruni, L. (2006) Civil Happiness: Economics and Human flourishing in historical perspective. New York:
Routledge.
Bruni, L. and Stanca, L. (2008) Watching Alone: Relational Goods, Television and Happiness. Journal of
Economic Behavior & Organization 65: 506–528.
Brunner, E. and Marmot, M. (1999) Social organization, stress, and health. In M.G. Marmot and R.G.
Wilkinson (eds), Social Determinants of Health (pp. 17–43). Oxford: Oxford University Press.
Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233
C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 229

Bukenya, J.O., Gebremedhin, T.G. and Schaeffer, P.V. (2003) Analysis of rural quality of life and health: a
spatial approach. Economic Development Quarterly 17: 280–293.
Cantarero, D., Pascual, M. and Sarabia, J.M. (2005) Effects of income inequality on population health: new
evidence from the European Community Household Panel. Applied Economics 35: 87–91.
Caviglia-Harris, J.L., Chambers, D. and Kahn, J.R. (2009) Taking the “U” out of Kuznets. A comprehensive
analysis of the EKC and environmental degradation. Ecological Economics 68: 1149–1159.
Chen, L. and Murray, C. (1992) Understanding morbidity change. Population and Development Review 18:
481–504.
Clark, A.E. and Oswald, A.J. (1994) Unhappiness and unemployment. Economic Journal 104: 648–659.
Clark, A.E., Frijters, P. and Shields, M.A. (2006) Income and happiness: evidence, explanations and economic
implications. Paris-Jourdan Sciences Economiques, Working Paper No. 2006–24, Paris.
Clark, A.E., Frijters, P. and Shields, M.A. (2008) Relative income, happiness and utility: an explanation for
the Easterlin paradox and other puzzles economic. Journal of Economic Literature 46: 95–144.
Cole, M.A., Rayner, A.J. and Bates, J.M. (1997) The environmental Kuznets curve: an empirical analysis.
Environment and Development Economics 2: 401–416.
Conservation Foundation (1992) State of the Environment. Washington, DC.
Cornia, G.A., Rosignoli, S. and Tiberti, L. (2008) Globalisation and health: impact pathways and recent
evidence WIDER Research Paper N. 2008/74. Helsinki: WIDER.
Crimmins, E.M. (2004) Trends in the health of the elderly. Annual Review of Public Health 25: 79–98.
Crimmins, E., Hayward, M., Hagedorn, A., Saito, Y. and Brouard, N. (2009) Change in disability-free life
expectancy for Americans 70 years old and older. Demography 46: 627–646.
Darity, W. and Goldsmith, A.H. (1996) Social psychology, unemployment and macroeconomics. Journal of
Economic Perspectives 10: 121–140.
Deacon, R.T. and Norman, C. (2006) Does the environmental Kuznets curve describe how individual countries
behave? Land Economics 82: 291–315.
Deaton, A. (2002) Health, inequality and economic development. Journal of Economic Literature 41: 113–158.
Deaton, A. (2008) Income, health, and well-being around the world: evidence from the Gallup World Poll.
Journal of Economic Perspectives 22: 53–72.
De Bruyn, S.M., Van den Bergh, J. and Opschoor, J.B. (1998) Economic growth and emissions: reconsidering
the empirical basis of environmental Kuznets curve. Ecological Economics 25: 161–175.
Deci, E.L. and Ryan, R.M. (1985) Intrinsic Motivation and Self-Determination in Human Behaviour. New
York: Plenum Press.
De Vogli, R., Mistry, R., Gnesotto, R. and Cornia, G.A. (2005) Has the relation between income inequality
and life expectancy disappeared? Evidence from Italy and top industrialised countries. Journal of
Epidemiology and Community Health 59: 158–162.
Diener, E. (1996) Traits can be powerful, but are not enough: lessons from subjective well-being. Journal of
Research in Personality 30: 389–399.
Dinda, S. (2004) Environmental Kuznets curve hypothesis: a survey. Ecological Economics 49: 431–455.
Di Tella, R. and MacCulloch, R.J. (2008) Gross national happiness as an answer to the Easterlin paradox?
Journal of Development Economics 86: 22–42.
Di Tella, R., MacCulloch, R.J. and Oswald, A.J. (2001) Preferences over inflation and unemployment:
evidence from surveys of happiness. American Economic Review 91: 335–41.
Dolan, P., Peasgood, T. and White, M. (2008) Do you really know what makes us happy? A review of the
economic literature on the factors associated with subjective well-being. Journal of Economic Psychology
29: 94–122.
Easterlin, R.A. (1974) Does economic growth improve the human lot? Some empirical evidence. In P.A.
David, W.R. Melvin (eds), Nations and Households in Economic Growth: Essays in Honour of Moses
Abramovitz (pp. 89–125). New York: Academic Press.
Easterlin, R.A. (2001) Income and happiness: towards a unified theory. Economic Journal 111: 465–484.
Easterlin, R.A. (2005) Building a better theory of well-being. In L. Bruni and P.L. Porta (eds), Economics
and Happiness. Framing the analysis (pp. 29–64). Oxford: Oxford University Press.
Easterlin, R.A. (2009) Lost in transition: life satisfaction on the road to capitalism. Journal of Economic
Behavior & Organization 71: 130–145.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
230 BORGHESI AND VERCELLI

Easterlin, R.A. and Angelescu, L. (2009) Happiness and growth the world over: time series evidence on the
happiness-income paradox. IZA Discussion Paper No. 4060. Bonn: Institute for the Study of Labor.
Easterlin, R.A. and Sawangfa, O. (2009) Happiness and economic growth: does the cross section predict
time trends? Evidence from developing countries. In E. Diener, J. Heliwell and D. Kahneman (eds),
International Differences in Well Being. Princeton: Princeton University Press.
Elstad, J.I., Dahl, E. and Hofoss, D. (2006) Associations between relative income and mortality in Norway:
a register-based study. European Journal of Public Health 16: 640–644.
Fahey, T. and Smyth, E. (2004) Do subjective indicators measure welfare? Evidence from 33 European
societies. European Societies 6: 5–27.
Ferrer-i-Carbonell, A. and Gowdy, J.M. (2007) Environmental degradation and happiness. Ecological
Economics 60: 509–516.
Fiscella, K and Franks, P. (1997) Poverty or income inequality as predictor of mortality: longitudinal cohort
studies. British Medical Journal 314: 1724–1728.
Frank, R.H. (1985) Choosing the Right Pond. New York: Oxford University Press.
Freedman, V.A., Crimmins, E.M., Schoeni, R.F., Spillman, B.C., Aykan, H., Kramarow, E., Land, K., Lubitz,
J., Manton, K.G., Martin, L.G., Shinberg, D. and Waidmann, T.A. (2004) Resolving inconsistencies in
trends in old-age disability: report from a technical working group. Demography 41: 417–441.
Frey, B.S. and Stutzer, A. (2002) Happiness and Economics. Princeton: Princeton University Press.
Gerdtham, U. and Johannesson, M. (2001) The relationship between happiness, health and socio-economic
factors: results based on Swedish micro data. Journal of Socio-Economics 30: 553–557.
Griffith, R. (2000) How important is business R&D for economic growth and should the government subsidise
it? IFS Briefing Note No. 12. London: Institute for Fiscal Studies.
Graham, C. (2008) Happiness and health: lessons—and questions—for public policy. Health Affairs 27:
72–87.
Graham, C. and Felton, A. (2005) Does inequality matter to individual welfare: an exploration based on
happiness surveys in Latin America, Center on Social and Economic Dynamics Working Papers Series
No. 38. Washington, DC: The Brookings Institution.
Gravelle, H. (1998) How much of the relation between population mortality and unequal distribution of
income is a statistical artefact? British Medical Journal 316: 382–385.
Gravelle, H. and Sutton, M. (2006) Income, relative income, and self-reported health in Britain 1979–2000.
CHE Research Paper 10, University of York. York: Centre for Health Economics.
Grossman, G.M. and Krueger, A.B. (1995) Economic growth and the environment. Quarterly Journal of
Economics 110: 353–377.
Gui, B. (1996) On relational goods: strategic implications of investment in relationships. International Journal
of Social Economics 23: 260–278.
Gui, B. and Sugden, R. (2005) Economics and Social Interaction: Accounting for Interpersonal Relations.
Cambridge: Cambridge University Press.
Hagerty, M.R. (2000) Social comparisons of income in one’s community: evidence from national surveys of
income and happiness. Journal of Personality and Social Psychology 78: 746–771.
Haller, M. and Hadler, M. (2006) How social relations and structures can produce happiness and unhappiness:
an international comparative analysis. Social Indicators Research 75: 169–216.
Hansen, K.S. and Østerdal, L.P. (2006) Models of quality-adjusted life years when health varies over time:
survey and analysis. Journal of Economic Surveys 20: 229–255.
Heliwell, J.F. (2003) How’s life? Combining individual and national variables to explain subjective well-
being. Economic Modelling 20: 331–360.
Hirsch, F. (1976) The Social Limits to Growth. Cambridge, MA: Harvard University Press.
Huppert, F. (2006) Positive emotions and cognition: developmental, neuroscience and health perspectives. In
J. Forges (ed.), Hearts and Minds: Affective Influences on Social Cognition and Behaviour. Philadelphia:
Psychology Press.
Judge, K. (1995) Income distribution and life expectancy: a critical appraisal. British Medical Journal 311:
1282–1285.
Kahneman, D. (1999) Objective happiness. In D. Kahneman, E. Diener and N. Schwartz (eds), Well-being:
Foundations of Hedonic Psychology. New York: Russel Sage Foundation.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 231

Kahneman, D. and Riis, J. (2005) Living, and thinking about it, two perspectives. In F.A. Huppert, B.
Kaverne and N. Baylis (eds), The Science of Well-being. Oxford: Oxford University Press.
Kaplan, G.A., Pamuk, E.R., Lynch, J.W., Cohen, R.D. and Balfour, J.L. (1996) Inequality in income and
mortality in the United States: analysis of mortality and potential pathways. British Medical Journal
312: 999–1003.
Kawachi, I. and Kennedy, BP. (1997) The relationship of income inequality to mortality - Does the choice
of indicator matter? Social Science and Medicine 45: 1121–1127.
Kawachi, I. and Kennedy B.P. (2002) The Health of Nations: Why Inequality Is Harmful to Your Health. New
York: The New Press.
Kawachi, I., Kennedy, B.P., Lochner, K. and Prothrow-Stith, D. (1997) Social capital, income inequality and
mortality. American Journal of Public Health 87: 1491–1498.
Kawachi, I., Kennedy, B.P. and Wilkinson, R.G. (1999) Income Inequality and Health. Vol. I. The Society and
Population Health Reader. New York: New Press.
Kivimäki, M., Virtanen, M., Vartia, M., Elovainio, M., Vahtera, J. and Keltikangas-Järvinen, L. (2003)
Workplace bullying and the risk of cardiovascular disease and depression. Occupational and
Environmental Medicine 60: 779–783.
Koivumaa-Honkanen, H., Honkanen, R., Viinamaeki, H., Heikkilae, K., Kaprio, J. and Koskenvuo, M. (2001)
Life satisfaction and suicide: a 20 year follow-up study. American Journal of Psychiatry, 158: 433–439.
Korpi, T. (1997) Is utility related to employment status? Employment, unemployment, labor market policies
and subjective well-being among Swedish youth. Labour Economics 4: 125–147.
Koskenvuo, M., Kaprio, J., Rose, R.J., Kesaniemi, A., Sarna, S., Heikkila, K. and Langinvainio, H. (1988)
Hostility as a risk factor for mortality and ischemic heart disease in men. Psychosomatic Medicine 50:
330–340.
Kouvonen, A., Oksanen, T., Vahtera, J., Stafford, M., Wilkinson, R., Schneider, J., Väänänen, A., Virtanen,
M., Cox, S.J., Pentti, J., Elovainio, M. and Kivimäki, M. (2008) Low workplace social capital as
a predictor of depression: the Finnish public sector study. American Journal of Epidemiology 167:
1143–1151.
Kunzli N., Kaiser, R., Medina, S., Studnicka, M., Chanel, O., Filliger, P., Herry, M., Horak, F. Jr.,
Puybonnieux-Texier, V., Quenel, P., Schneider, J., Seethaler, R., Vergnaud, J.C. and Sommer, H. (2000)
Public-health impact of outdoor and traffic-related air pollution: a European assessment. Lancet 356:
795–801.
Layard, R. (2006) Happiness: Lessons from a New Science. London: Penguin.
Leigh, A. and Jencks, C. (2007) Inequality and mortality: long run evidence from a panel of countries.
Journal of Health Economics 26: 1–24.
Lelkes, O. (2006) Knowing what is good for you. Empirical analysis of personal preferences and the
“objective good”. Journal of Socio-Economics, 35: 285–307.
Likken, D. and Tellegen, A. (1996) Happiness is a stochastic phenomenon. Psychological Science 7: 180–189.
Lucas, R. (1981) Studies in Business-Cycle Theory. Cambridge, MA: MIT Press.
Lucas, R.E., Clark, A., Georgellis, Y. and Diener, E. (2004) Unemployment alters the set point for life
satisfaction. Psychological Science 15: 8–13.
Lynch, J., Kaplan, G.A., Pamuk, E.R., Cohen, R.D., Heck, K.H., Balfour, J.L. and Yen, I.H. (1998) Income
inequality and mortality in metropolitan areas of United States. American Journal of Public Health, 88:
1074–1080.
Lynch, J., Smith, G.D., Kaplan, G.A. and House, J.S. (2000) Income inequality and mortality: importance to
health of individual income, psychosocial environment, or material conditions. British Medical Journal,
320: 1200–1204.
Manton, K.G., Gu, X. and Lamb, V.L. (2006) Change in chronic disability from 1982 to 2004/2005 as
measured by long-term changes in function and health in the U.S. elderly population. Proceedings of
the National Academy of Sciences of the United States of America 103: 18374–18379.
Marmot, M. (2004) Status Syndrome. New York: Henry Holt.
Ng, Y.K. (1978) Economic growth and social welfare: the need for a complete study of happiness. Kyklos
31: 575–587.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
232 BORGHESI AND VERCELLI

Nussbaum, M. (1986) The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Ethics. Cambridge:
Cambridge University Press.
O’Connell, M. (2004) Fairly satisfied: economic equality, wealth and satisfaction. Journal of Economic
Psychology 25: 297–305.
OECD – Organisation for Economic Cooperation and Development. (2001) Health at a Glance. Paris: OECD.
OECD – Organisation for Economic Cooperation and Development. (2005) Society at a Glance: OECD
Social Indicators 2005. Paris: OECD.
Palmore, E. (1969) Predicting longevity: a follow-up controlling for age. Journal of Gerontology, 39: 109–116.
Pichler, F. (2006) Subjective quality of life of young Europeans. Feeling happy but who knows why? Social
Indicators Research, 75: 419–444.
Pope, A.M. and Tarlov, A. (1991) Disability in America: Toward a National Agenda for Prevention.
Washington, DC: National Academy Press.
Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K. and Thurston, G.D. (2002) Lung
cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. Journal of
the American Medical Association 287: 1132–1141.
Preston, S.H. (1975) The changing relation between mortality and level of economic development. Population
Studies 29: 231–248.
Pugno, M. (2007) The subjective well-being paradox: a suggested solution based on relational goods. In L.
Bruni and P.L. Porta (eds.), Handbook on the Economics of Happiness (pp. 263–289). London: Edward
Elgar.
Pulkki, L., Kivimäki, M., Elovainio, M., Viikari, J. and Keltikangas-Järvinen, L. (2003) Contribution of
socioeconomic status to the association between hostility and cardiovascular risk behaviors: a prospective
cohort study. American Journal of Epidemiology 158: 736–742.
Putnam, R. (2000) Bowling Alone: the Collapse and Revival of American Community. New York: Simon and
Schuster.
Romer, P. (1990) Endogenous technological change. Journal of Political Economy 98: 71–102.
Roca, J., Padilla, E., Farré, M. and Galletto, V. (2001) Economic growth and atmospheric pollution in Spain:
discussing the environmental Kuznets curve hypothesis. Ecological Economics 39: 85–99.
Ryff, C.D. and Singer, B. (2000) Interpersonal flourishing: a positive health for the new millennium.
Personality and Social Psychology Review 4: 30–44.
Sales, S.M. and House, J. (1971) Job dissatisfaction as a possible risk factor in coronary heart disease.
Journal of Chronic Diseases, 23: 861–873.
Sapolsky, R.M. (1998) Why Zebras Don’t Get Ulcers. A Guide to Stress, Stress-Related Disease and Coping,
2nd edn. New York: W.H. Freeman.
Saunders, P.A. (1996) Income and Welfare: Special Article - Poverty and Deprivation in Australia. In Year
Book Australia 1996, Commonwealth of Australia, Canberra.
Scitovsky, T. (1976) The Joyless Economy: An Inquiry into Human Satisfaction and Dissatisfaction. Oxford:
Oxford University Press.
Sen, A. (2002) Health: perception versus observation. British Medical Journal 324: 860–861.
Selden, T.M. and Song, D. (1994) Environmental quality and development: is there a Kuznets curve for air
pollution emissions? Journal of Environmental Economics and Management 27: 147–162.
Shafik, N. (1994) Economic development and environmental quality: an econometric analysis. Oxford
Economic Papers 46: 757–773.
Shen, J. (2006) A simultaneous estimation of environmental Kuznets curve: evidence from China. China
Economic Review 17: 383–394.
Stern, D.I., Common, M.S. and Barbier, E.B. (1996) Economic growth and environmental degradation: the
environmental Kuznets curve and sustainable development. World Development 24: 1151– 1160.
Stevenson, B. and Wolfers, J. (2008) Economic growth and subjective well-being: reassesing the Easterlin
paradox. IZA Discussion Paper No. 3654, Institute for the Study of Labor, Bonn, Germany.
Subramanian, S.V., Kim, D.J. and Kawachi, I. (2002) Social trust and self-rated health in US communities:
a multilevel analysis. Journal of Urban Health 79: S21–S34.
Swift, J. (1726) Gulliver’s Travels. ed. 2005. Oxford: Oxford University Press.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd
HAPPINESS AND HEALTH: TWO PARADOXES 233

Tarlov, A.R. and St. Peter, R.F. (2000) The Society and Population Health Reader, Volume II, A State and
Community Perspective. New York: The New Press.
UNDP - United Nations Development Programme (2007) Human Development Report 2007/2008.
Basingstoke, UK: Macmillan.
United Nations – Department of Economic and Social Affairs. (2005) World Population Prospects. The 2004
Revision. New York: United Nations.
Uslaner, E. (2001) The moral foundations of trust. Cambridge: Cambridge University Press.
Veenhoven, R. (1984) Conditions of Happiness. Dordrecht: Kluwer Academic.
Vercelli, A. (1991) Methodological Foundations of Macroeconomics. Keynes and Lucas. Cambridge:
Cambridge University Press.
Vincent, J.R. (1997) Testing for environmental Kuznets curves within a developing country. Environment
and Development Economics 2: 417–431.
Wagstaff, A. and van Doorslaer, E. (2000) Income inequality and health: what does the literature tell us?
Annual Review of Public Health 21: 543–567.
Waidmann, T.A. and Liu, K. (2000) Disability trends among elderly persons and implications for the future.
Journal of Gerontology, 55B: S298–S307.
Waidmann, T.A. and Manton, K.G. (1998) International Evidence on Disability Trends among the Elderly.
Washington, DC: U.S. Department of Health and Human Services.
Warr, P. (1999) Well-being and the workplace. In D. Kahneman, E. Diener, and N. Schwarz (eds.), The
Foundations of Hedonic Psychology (pp. 392–412). New York: Russel Sage Foundation.
WHO – World Health Organization. (1997) Health and Environment in Sustainable Development: Five Years
after the Earth Summit. Geneva: WHO.
WHO – World Health Organization. (2004) World Health Report 2004. Geneva: WHO.
WHO – World Health Organization. (2005) Preventing Disease Through Healthy Environments: Towards an
Estimate of the environmental Burden of Disease. Geneva: WHO.
Wilkinson, R.G. (1992) Income distribution and life expectancy. British Medical Journal 304: 165–168.
Wilkinson, R.G. (2002) Socioeconomic status and health. In E. Ziglio, L. Levi and E. Bath (eds), Investment
for Health: A Discussion of the Role of Economic and Social Determinants. Copenhagen: World Health
Organization.
Wilkinson, R.G., Kawachi, I. and Kennedy, B.P. (2008) Mortality, the social environment, crime and violence.
Sociology of Health and Illness 20: 578–597.
Williams, R.B., Feaganes, J. and Barefoot, J.C. (1995) Hostility and death rates in 10 U.S. cities.
Psychosomatic Medicine 57: 94.
Winkelmann, L. and Winkelmann, R. (1998) Why are the unemployed so unhappy? Evidence from panel
data. Economica 65: 1-15.
World Bank. (1992) World Development Report 1992. New York: Oxford University Press.
World Bank. (2008) World Development Indicators. Washington, DC: The World Bank.

Journal of Economic Surveys (2012) Vol. 26, No. 2, pp. 203–233


C 2010 Blackwell Publishing Ltd

You might also like