The mystery of the U-shaped relationship between happiness and age

https://doi.org/10.1016/j.jebo.2012.03.008Get rights and content

Abstract

In this paper, we address the puzzle of the relationship between age and happiness. Whilst the majority of psychologists have concluded there is not much of a relationship at all, the economic literature has unearthed a possible U-shape relationship with the minimum level of satisfaction occurring in middle age (35–50). In this paper, we look for a U-shape in three panel data sets, the German Socioeconomic Panel (GSOEP), the British Household Panel Survey (BHPS) and the Household Income Labour Dynamics Australia (HILDA). We find that the raw data mainly supports a wave-like shape that only weakly looks U-shaped for the 20–60 age range. That weak U-shape in middle age becomes more pronounced when allowing for socio-economic variables. When we then take account of selection effects via fixed-effects, however, the dominant age-effect in all three panels is a strong happiness increase around the age of 60 followed by a major decline after 75, with the U-shape in middle age disappearing such that there is almost no change in happiness between the age of 20 and 50.

Introduction

What is the relationship between happiness and age? Do we become more miserable as we age, or, is our happiness relatively constant throughout our lives with only the occasional special event (marriage, birth, promotion, and illness) temporarily raising or reducing our happiness, or do we actually get happier as we age?

The answer to this question in the recent economic literature is that the age–happiness relationship is U-shaped.1 This finding holds for the US, Germany, Britain, Australia, Europe, and South Africa. The stylised finding is that individuals gradually become unhappier after their 18th birthday, with a minimum around 50, followed by a gradual upturn in old age. The predicted effect of age can be quite large. For example, the predicted difference in average happiness between an 18 year old and a 50 year old from regressions can be as much as 1.5 points on a 10-point-scale.

This recent economics literature, however, conflicts with an old psychology literature that finds no happiness-age relationship (Cantril, 1965). Palmore and Luikart (1972) comment in their review; ‘Several variables thought to be related to life satisfaction had little or no relationship: age, sex, total social contacts’. More recently, the psychologists Dear et al. (2002) postulate a slight reduction in life satisfaction as people age, due to the prevalence of high life satisfaction becoming less common at higher ages. From this reading, it is clear that either the psychologists have overlooked something important for a long time or that the methodology of economists begets different answers. This paper intends to find out, which it is.

We re-examine the age–happiness relationship and delve into the methodological aspects to provide an explanation for the difference of opinion between economists and psychologists. We essentially want to know if the U-shape that economic scholars find is an artefact or real, and what the actual relationship between age and life satisfaction is. We re-examine the age–happiness relationship in three often-used panel datasets, the German Socio Economic Panel (the GSOEP), the British Household Panel Survey (BHPS), and the Household Income Labour Dynamics Australia (HILDA), which all have an extensive set of variables on the individual level. This data-richness allows us to not only replicate the findings of other studies based on cross-sectional data, but, furthermore, allows us to explore the dynamic interplay between age, covariates, unobserved heterogeneity, and happiness.

The format of this paper is to let the solution to the puzzle of the age–happiness relationship progressively unfold. We first briefly review the recent literature where we summarise the main findings of others, as well as their methodology. Then we present the data we have and show that we can indeed replicate a U-shape in happiness when we run similar regressions to those in the literature. We then go through a succession of reasons for both the raw relationship between happiness and age in these panels, as well as the changes in coefficients of age-related variables as more factors are included. This includes the possibility: that the age–happiness relationship is dominated by a happiness reduction found in early adulthood (age 18–22); that found age effects are due to estimation biases arising from selectivity, or; that it is a truly robust finding. We find that selection, i.e. fixed effects, and are extremely important for the age–happiness puzzle. Not only does the inclusion of fixed-effects change the coefficients of important age-varying factors (such as employment and income), which in turn changes the found residual effects of age directly, but it also turns out that the raw relation is heavily tainted by selection effects; the panels seem to over-sample particularly happy very old individual and particularly unhappy middle age individuals, leading these datasets to exaggerate the happiness decline in middle-age and to underestimate the decline in very old age.

Section snippets

Literature review

Whilst a lot of the economic literature on the age–happiness relationship is recent, there have been earlier discussions of it (see Theodossiou, 1998 for a discussion of the history of this issue). Until the early 2000s, the opinion of economists about the effect of age was still divided. Clark and Oswald (1994) found a U-shaped pattern for the UK, whilst Winkelmann and Winkelmann (1998) found no U-shape in happiness but simply a very strong negative effect of age. Easterlin et al. (1993) using

The GSOEP

We use the 1984–2002 waves of the German Socio-Economic Panel (GSOEP, 2008), a representative 18-year panel of the German population. The first wave (1980) included only the Federal Republic of Germany; it has included the former East Germany since 1990. We use only the information on West Germany in order to be able to abstract from the importance of the 1990 German reunification, which had a tremendous impact on the lives and satisfaction levels of East Germans (Frijters et al., 2004). The

Is there a U-shape in the raw data?

For all analyses that follow the full regression tables are shown in Appendix B, but we tell the story using graphs and summary tables in the main text. We experimented using both simple least squares (which is the dominant method in the literature) and latent-variable analyses (for cross-sectional as well as fixed-effects analyses) but we found, as in Ferrer-i-Carbonell and Frijters (2004), that there is no qualitative difference, so we choose to present the least squares results here whilst

The relation is due to the very young and the very old

A naive first-thought is that there is a particular issue with the early ages, i.e. age 18–22, and with high ages, i.e. those above 80. This is because the happiness decline is particularly steep for the early years and one may worry about the selectivity of those who are still alive at very high ages; they could be much happier or much less happy than others. This makes one wonder if the young are being overly optimistic about their actual levels of happiness and that the happiness of the very

The actual age–happiness relationship

In order to arrive at a consolidated model of the age–happiness relation, we must allow for a more flexible age-profile, age-specific selectivity, and panel-dependent response profile than hitherto. We wish to answer two questions9:

  • 1.

    How does raw happiness vary over the life cycle for a given individual with a given level of initial happiness?

  • 2.

    What is the potential causal impact of age on happiness ceteris

Robustness analyses

We here briefly mention the robustness analyses we ran. We redid everything with latent-variable techniques rather than linear regressions. To this end we used ordered logits as a cross-sectional model and the recent BUC estimator from Baetschmann et al. (2011), which is a fixed-effect conditional logit estimator. The results are in Appendix C, Table C1, Table C2, Table C3. As in the main text above, the highly significant and positive effect on age-squared found in the cross-section

Conclusions and discussion

This paper started out with the puzzling findings of other researchers of a U-shaped relationship between age and happiness. We replicated this relationship for Germany, Australia, and Britain using well-known panel datasets, the GSOEP, the HILDA, and the BHPS. In all three cases, the age–happiness profile became a much clearer U-shape when adding commonly used socio-economic variables. This emergence of the U-shape was not dependent on the inclusion of individuals aged 18–22 or those above 80.

References (54)

  • J.V. Terza

    Estimating linear models with ordinal qualitative regressors

    Journal of Econometrics

    (1987)
  • I. Theodossiou

    The effects of low-pay and unemployment on psychological well-being: a logistic regression approach

    Journal of Health Economics

    (1998)
  • D.H. Wolpert

    Why income comparison is rational

    Games and Economic Behavior

    (2010)
  • M. Argyle et al.

    Causes and correlates of happiness

    Well-Being: The Foundations of Hedonic Psychology

    (1999)
  • G. Baetschmann et al.

    Consistent Estimation of the Fixed Effects Ordered Logit Model

    (2011)
  • Beatton, T., Frijters, P., 2009. Do Changes in the Lives of Our Peers Make Us Unhappy? Working Paper Series. Queensland...
  • D. Bell et al.

    The Scots may be brave but they are neither healthy nor happy

    Scottish Journal of Political Economy

    (2007)
  • BHPS

    British Household Panel Survey

    (2010)
  • D.G. Blanchflower

    International Evidence on Well-Being

    (2008)
  • D.G. Blanchflower et al.

    Well-Being Over Time in Britain and the USA

    (2001)
  • D.G. Blanchflower et al.

    Is Well-being U-Shaped Over the Life Cycle?

    (2007)
  • D.G. Blanchflower et al.

    The U-Shape without Controls

    (2009)
  • H. Cantril

    The Pattern of Human Concerns

    (1965)
  • Clark, A.E., 2006. Born To Be Mild? Cohort Effects Don’t Explain Why Well-Being is U-shaped in Age (vol. Working Paper...
  • A.E. Clark et al.

    Relative income, happiness, and utility: an explanation for the Easterlin paradox and other puzzles

    Journal of Economic Literature

    (2008)
  • A.E. Clark et al.

    Unhappiness an unemployment

    Economic Journal

    (1994)
  • P.T. Costa et al.

    Influence of extraversion and neuroticism on subjective well-being: happy and unhappy people

    Journal of Personality and Social Psychology

    (1980)
  • Cited by (0)

    We would like to thank conference, staff, and seminar participants for useful comments and suggestions. The usual disclaimer applies.

    View full text