Transcript Life Cycle Happiness and Its Sources

Life Cycle Happiness
and Its Sources
Intersections of Demography,
Economics and Psychology
(May, 2007)
Richard A. Easterlin
University of Southern California
Questions
1. On average, at what stage of life are
people happiest?
– On the threshold of their adult lives?
– At mid-life?
– In their golden years?
2. What factors principally determine the life
cycle pattern of happiness?
– personality and genetics?
– Life circumstances (work, family,
health, aspirations etc.)?
2
Part I. Background (30 slides)
Part II. Life Cycle Happiness (17 slides)
A. Nature
B. Causes
Part III. Conclusions and Implications
(10 slides)
3
PART I. BACKGROUND
How is happiness measured?
Present study:
Taken all together, how would you say things
are these days - would you say that you
are very happy, pretty happy, or not too
happy?
4
How measured? Variations:
Scaling: 0 to 10 ladder, etc.
Related concepts:
• General life satisfaction
• Positive Affect
• Negative Affect
Single vs. multiple item constructs
5
Can We Believe Self-Reports of
Well-Being? (1)
Economics: No. Can’t trust what people say.
“Economists, as a rule, are not concerned with the
internal thought processes of the decision maker
or in the rationalizations that the decision maker
offers to explain his or her behavior. Economists
believe that what people do is more relevant
than what they say" (Fuchs, 1983).
6
Can We Believe Self-Reports of
Well-Being? (2)
Economics: No.
“[A] large experimental literature by and large
supports economists’ skepticism of subjective
questions [and] … cast[s] serious doubts on
attempts to use subjective data as dependent
variables….”. (Bertrand and Mullainathan,
2001)
But see: Jonathan M. Gruber and Sendhil
Mullainathan (2005) “Do Cigarettes Make
Smokers Happier”, Advances in Economic
Analysis & policy: Vol. 5: No. 1.
7
Economics’ View of Self-Reports
8
Can We Believe Self-Reports of
Well-Being? (3)
Psychology: Yes. Not perfect, but
meaningful.
1. Reliability: test – retest studies
9
Can We Believe Self-Reports of
Well-Being? (4)
2. Validity: consistency with
- external observer ratings (spouse,
family, friends, professional therapist)
- physical, neurological evidence (facial
expressions, brain waves, stress
hormones)
- theoretically related evidence
(psychological counseling, symptoms
of anxiety and worry (headaches,
sleeplessness, suicide attempts))
10
Can We Believe Self-Reports of
Well-Being? (5)
3. Comparability: the things people say are
important for their happiness (living
conditions, family, health, work, personal
traits)
11
Why Study Happiness?
Happiness measures embody a different
conception of the nature of well-being
than that commonly accepted in
economics
Economics: Objective Well-being
Psychology: Subjective Well-being
12
Economics: Objective Well-being (OWB)
Revealed preference:
Well-being is evidenced by a person’s
choices and can safely be inferred from
objective measures such as household
income or, more generally, GDP per
capita.
Example: Growth of GDP per capita in
Japan, China.
13
Psychology: Subjective Well-being (SWB)
The effect on well-being of objective
circumstances is mediated by
psychological mechanisms such as
• Hedonic Adaptation (Brickman et al,
1972)
• Social Comparison
14
Psychology: Skepticism of OWB
“I cannot feel satisfied that the correspondence
between such objective measures as amount of
money earned, number of rooms occupied, or
type of job held, and the subjective satisfaction
with these conditions of life, is close enough to
warrant accepting the one as replacement for
the other”
(psychologist Angus Campbell, 1972)
15
Does It Matter (1)?
GDP per Capita and Mean Satisfaction,
Japan, 1958 - 1987
Satisfaction
GDP per capita
300
100
GDP per capita
(1958 = 100)
500
Source: Easterlin (1995)
16
Does It Matter (2)?
GDP per Capita, Per Cent Satisfied, and Per
Cent Dissatisfied, China, 1994 - 2004
90
300
80
250
60
200
50
GDP
40
150
Dissatisfied
30
20
GDP per capita
(1994 = 100)
% satisfied or dissatisfied
Satisfied
70
100
10
0
50
1994
1997
1999
2004
Year
Source: Kahneman and Krueger (2006)
17
Economics’ View of Well-Being
18
But if people adapt to circumstances,
as psychologists assert, then what
determines SWB?
The Setpoint Model:
Individual happiness fluctuates
around a setpoint given by:
personality, genetic heritage
19
The setpoint model:
“No one doubts that making the team, being
promoted at work, or winning the lottery tends
to bring about an increment in happiness, just
as flunking out, being laid off, or a disastrous
investment would be likely to diminish one’s
feelings of well-being…. [H]owever, the
effects of these events appear to be transitory
fluctuations about a stable temperamental set
point or trait that is characteristic of the
individual.”
(Lykken and Tellegen, 1996)
20
Well-being in the Setpoint Model
21
The setpoint model: Are you kidding?(1)
“Objective life circumstances have a
negligible role to play in a theory of
happiness.”
Richard Kammann (1983)
“Objective circumstances appear to be limited
in the magnitude, scope, and particularly
duration of their effects on psychological wellbeing, which, in the long run, is likely to reflect
instead stable characteristics of the
individual.”
(Costa et al, 1987)
22
The setpoint model: Are you kidding?(2)
“Each individual may be on a personal treadmill
that tends to restore well-being to a
predetermined setpoint after each change of
circumstances.”
(Kahneman, 1999)
”Our human capacity for adaptation… helps
explain a major conclusion of subjective wellbeing research, as expressed by the late Richard
Kammann (1983): ‘Objective life circumstances
have a negligible role to play in a theory of
happiness’”
(Myers 2000, p.60)
23
The setpoint model: Are you kidding?(3)
Chance events like personal tragedies, illness, or
sudden strokes of good fortune may drastically
affect the level of happiness, but apparently these
effects do not last long.”
(Csikszentmihalyi and Hunter, 2003)
24
The setpoint model:
Representative? (1)
“The assumption that happiness
setpoints exist has guided much of the
current theory and research on SWB…”
(Lucas et al 2004, p.8)
25
The setpoint model:
Representative? (2)
Media Reactions
o Time (The Science of Happiness, Jan. 17,2005)
o “Circumstances don’t seem to have much
effect on happiness” (Kahneman, p.A32)
o NY Times Magazine Section, Sept. 7, 2003
o “The Futile Pursuit of Happiness” by John
Gerstner (based on interviews with psychologists
Gilbert, Wilson, Loewenstein, Kahneman)
26
The setpoint model:
Representative? (3)
Policy (1): The “Iron Law of Happiness”
“The influence of genetics and personality
suggests a limit on the degree to which
policy can increase SWB… Changes in the
environment, although important for shortterm well-being, lose salience over time
through processes of adaptation, and have
small effects on long-term SWB.”
(Diener and Lucas, 1999)
27
The setpoint model:
Representative? (4)
Policy (2): The “Iron Law of Happiness”
“According to the hedonic treadmill model, good and
bad events temporarily affect happiness, but people
quickly adapt back to hedonic neutrality. The theory,
which has gained widespread acceptance in recent
years, implies that individual and societal efforts to
increase happiness are doomed to failure.”
(Diener, E., Lucas, R. E., & Scollon, C. N. (2006). Beyond
the hedonic treadmill: Revisions to the adaptation theory
of well-being. American Psychologist, 61, 305-314)
28
The setpoint model: Becoming less
representative? (1)
1. Misread of Brickman et al, 1972:
quadraplegics and paraplegics
significantly less happy.
2. Adaptation is not always complete
(noise, cosmetic surgery) Frederick and
Loewenstein, 1999.
29
The setpoint model: Becoming less
representative? (2)
3. Retreat on the policy front: Diener and
Seligman, 2004, proposal for
governmental measurement of SWB.
4. Retreat on complete adaptation: Diener,
Lucas, Scollon, 2006. “Happiness can
and does change” (p. 309)
30
Implications for
Economics View of Well-Being
Economics is not allowing for any subjective
processing of objective circumstances
How sizeable are hedonic adaptation and social
comparison? Enough to invalidate inferences
based on objective measures?
Perhaps we ought, at least, to look at data on
Happiness
31
Recent Research on
Economics of Happiness
Books:
L. Bruni Civil Happiness (2006)
R. Layard, Happiness: Lessons from a New Science
(2005)
L. Bruni and P. Porta,eds. Economics and Happiness
(2005)
B.M.S. van Praag and A. Ferrer-i-Carbonell, Happiness
Quantified (2004)
C. Graham and S. Petinato, Happiness and Hardship
(2002)
R.A. Easterlin, ed. Happiness in Economics (2002)
B. Frey and A. Stutzer, Happiness and Economics (2002)
32
Recent Research on
Economics of Happiness
Symposia, Surveys:
Journal of Economic Perspectives, 2006
(Kahneman and Krueger, DiTella and
MacCuuloch)
Journal of Economic Literature, 2002
(Frey and Stutzer)
Journal of Economic Behavior and
Organization, 2001 (five articles)
Economic Journal, 1997
(Oswald, Frank)
New Palgrave, forthcoming
(Carol Graham)
33
PART II. Life Cycle Happiness
1. On average, at what stage of life are people
happiest?
– On the threshold of their adult lives?
– At mid-life?
– In their golden years?
2. What factors principally determine the life cycle
pattern of happiness?
– personality and genetics?
– Life circumstances (work, family, health,
aspirations etc.)?
34
Life Cycle Happiness:
What do we know?
(psychologists in italics)
Constant: David Myers (2000)
Costa et al (1987)
Easterlin and Schaeffer (1999)
Rising:
Argyle (2001)
Carstensen et al (2000)
Mroczek and Kolarz (1998)
Charles et al (2001)
35
What Do We Know?
(psychologists in italics)
Inverted U: Mroczek and Shapiro (2004)
U-shaped: Blanchflower and Oswald (2004)
Frey and Stutzer (2002), etc.
36
Nature of Life Cycle Happiness:
Present Study – Source and Measure
Data: US General Social Survey, 1973-1994
HAPPY: Global happiness
(scaled from 3 down to 1)
(one among many questions)
37
Demography
Synthetic Panel Method: The Ideal
Record the Reported Happiness of a
random sample of persons born in a given
year (birth cohort) as they age year by
year from 18 to 89.
Problem: Requires 71 years of data
38
Synthetic Panel Method: Actual
Piece together the full life course from
segments of experience of
single year birth cohorts (n=23,1119)
52 cohorts, born 1905-55
21 yrs. of data for each
42 cohorts, born 1884-1904 and 19561976, average 10.5 yrs. of data each
39
Method of estimating Life Cycle Pattern
(Ordered Logit Regression)
Dependent Variable: Happy
Independent Variables (5):
Age, with controls for
Year of Birth Race
Gender
Education
40
Life Cycle Happiness
2.4
Mean Happy
2.3
2.2
2.1
2.0
18
28
38
48
58
68
78
88
Age
41
Causes of Life Cycle Happiness:
Conceptual Framework
1. SWB is the net result of Satisfaction
with various Life Domains (goods,
family, work, health, etc.)
2. Domain Satisfaction depends on the
extent to which
Attainments (objective conditions)
match
Goals, Aspirations, Needs
(subjective norms)
42
Causes of Life Cycle Happiness:
Conceptual Framework: Antecedents
Psychology:
Angus Campbell et al (domain
satisfaction)
Social Indicators:
Alex Michalos (multiple
discrepancy theory)
Economics:
Simon and March (aspirations and attainments)
de la Croix (norm achievement model)
Duesenberry, Modigliani, Ng, Pollak (relative
income, habit formation)
Kahneman and Tversky (adaptation)
43
What domains?: What people say is
important for their happiness
44
Causes: Measures of Domain
Satisfaction (4) in Present Study
SATFIN: financial situation (3 to 1)
SATJOB: work including keeping house (4 to 1)
SATFAM: family life (7 to 1)
SATHEALTH: health (7 to 1)
45
Causes: Analytical steps (3)
1. Each domain: Estimate life cycle pattern
(same method as for Happy)
2. From individual data: Estimate Happy as
function of 4 domains
3. Predict Life Cycle Happy from Life Cycle
Domain Patterns obtained in step 1,
using regression of step 2.
46
Causes: step 1: Life Cycle
Happiness and Domain Satisfaction
2.4
2.3
Happy
2.2
2.1
2.0
2.5
6.1
2.4
5.9
SATFAM
(scale-->)
2.3
5.7
2.2
5.5
2.1
5.3
SATFIN
(<--scale)
2.0
5.1
1.9
4.9
18
28
38
48
58
Age
68
78
88
47
Step 1 (cont): Life Cycle
Happiness and Domain Satisfaction
2.4
2.3
Happy
2.2
2.1
2.0
3.6
6.1
SATHEALTH
(scale-->)
3.5
5.9
3.4
5.7
3.3
5.5
3.2
5.3
SATJOB
(<--scale)
3.1
5.1
3.0
4.9
18
28
38
48
58
Age
68
78
88
48
Causes: step 2: How Happiness
Varies with Domain Satisfaction
(Ordered Logit Regression)
Each domain has significant positive effect
on Happy.
Order of importance: family, financial, work,
health
49
Causes: step 3: Life Cycle
Happiness, Actual and Predicted
2.4
Mean Happy
2.3
Predicted
2.2
Actual
2.1
2.0
18
28
38
48
58
Age
68
78
88
50
PART III. Conclusion (1): Nature
1. At what stage of life are people
happiest?
Answer: On average, Happiness rises
slightly to midlife, and then declines
somewhat.
51
Conclusion (2): Causes
2.
What factors principally determine the life cycle
pattern of happiness?
Answer:
a. Up to midlife, increasing SATFAM and
SATJOB outweigh diminishing SATHEALTH.
b. Beyond midlife, decreasing SATFAM, SATJOB,
SATHEALTH outweigh increasing SATFIN.
52
If happiness is the outcome of
satisfaction in various domains,
what principally determines each
domain pattern of satisfaction –
objective conditions (economics)
or subjective conditions
(psychology)?
53
Satisfaction with Health:
Hypothetical possibilities
Complete Adaptation
Mean satisfaction
Mean satisfaction
“Objective” dominates
18 23 28 33 38 43 48 53 58 63 68 73 78 83 88
Age
18 23 28 33 38 43 48 53 58 63 68 73 78 83 88
Age
54
Health and Work Domain
Satisfaction:
2.4
2.3
Happy
2.2
2.1
2.0
3.6
6.1
SATHEALTH
(scale-->)
3.5
5.9
3.4
5.7
3.3
5.5
3.2
5.3
SATJOB
(<--scale)
3.1
5.1
3.0
4.9
18
28
38
48
58
Age
68
78
88
55
Family and Finances Domain
Satisfaction:
2.4
2.3
Happy
2.2
2.1
2.0
2.5
6.1
2.4
5.9
SATFAM
(scale-->)
2.3
5.7
2.2
5.5
2.1
5.3
SATFIN
(<--scale)
2.0
5.1
1.9
4.9
18
28
38
48
58
Age
68
78
88
56
Conclusion (3): Explaining
domain satisfaction
Objective conditions dominant:
sathealth, satjob, satfam
Subjective conditions dominant:
satfin
57
Implications (1)
For Measurement
The life cycle happiness and domain
satisfaction data come together
in a plausible way that supports the
view that self- reports on feelings
are meaningful.
58
Implications (2)
For Theory
Life cycle happiness: Findings support
model in which Life Cycle Happiness is
determined by satisfaction with major life
domains.
2. Domain satisfaction: Both objective and
subjective factors are important.
1.
Objective (SATFAM, SATHEALTH, SATJOB).
Subjective (SATFIN)
3.
To explain SWB a model building on both
economics and psychology is needed.
59
Implications (3)
For policy
There is no “iron law of happiness”
Example: Public policy increases Health
Better Health increases
SATHEALTH, SATFAM.
Increases in SATHEALTH,
SATFAM raise HAPPY
60
P.S. (for skeptics)
61
Envoi (for economists)
They seek it here; they seek it there
Economists seek it – everywhere
Is it GDP? SWB?
That elusive thing – UTILITY!
(apologies to Baroness Orczy)
62
THANK YOU
63