Study Horner (2014): study ZZ EU 14 2004
- Public
- 50-70 aged men, 16 western nations, 2004-2006
- Survey name
- Multiple surveys pooled
- Sample
- Respondents
- N = 18345
- Non Response
- Assessment
-
Interview: Computer Assisted Personal Interview (CAPI)
face-to-face and computer-aided personal interview
Correlate
- Authors's Label
- Retirement status
- Our Classification
-
-
- Related specification variables
-
-
- Operationalization
- Selfreport of current employment status
1. Fully retired
0. Other
a. Employed
b. Self-employed
c. Unemployed and looking for work
d. Permanently sick or disabled
e. Homemaker
f. Other (rentier, living off own property, student, doing voluntary work)
Observed Relation with Happiness
Weighted by country population, controlled for
- country fixed-effect
- married
- age, age squared, age cubed
- season of interview
Unaffected by additionally control for
- period to early retirement age
- if younger than cutoff
- if older than cutoff
- period to full retirement age
- if younger than cutoff
- if older than cutoff
- reached early retirement age
- reached full retirement age
Similar for analyses
- without Ireland samples
- without USA sample
- 40-60 aged men
- 60-80 aged men
Somewhat stronger when additionally controlled for
- period to early retirement age
- if younger than cutoff
- if older than cutoff
- period to full retirement age
- if younger than cutoff
- if older than cutoff
Interpretation: b (OLS) underestimates the effect of retirement because unhappy people tend to retire earlier, whereas b-iv (Instumental Variable Analysis) reveals a positive effect.
- below age 65: b-iv = +.28 (05)
- above age 65: b-iv = +2.6 (001)