Job Satisfaction, Well-Being, and Labour Mobility Francis Green
A little history of “job satisfaction” in Economics • Original prime justification: predictor of voluntary mobility (quits) (e.g. Freeman, 1978) • Data availability • Potential proxy for economics’ central concept “utility” • The job satisfaction research programme: estimating the determinants of job satisfaction 2
Refinements • What does it mean? An indicator unsure of its mother concept; a focus on the relationship of job satisfaction to decision utility (e.g. Levy-Garboua and Montmarquette, 2004, 2007) • What are the comparators behind comparison income and job satisfaction (e.g. Clark & Oswald, 1996) • Which domain best predicts quits (e.g. Clark, 2001)? • What are the distributional trends? e.g. Hammermesh, 2001; Green, 2006) • BUT, Some false directions: – equating job satisfaction directly to job quality or utility, comparing across cultures (Kristensen & Johansson, 2008). 3
Subjective Well-being • Arousal and pleasure-pain emotional space – Two axes commonly measured: • Enthusiasm-Depression & Contentment-Anxiety
• Job satisfaction seen as sitting along the pleasure-pain dimension, so incomplete • Perhaps well-being indicators might be preferable measures of welfare also within economics? 4
Evidence? – health effects – performance effects • Well-being and absenteeism; E-D axis more closely associated than C-A axis • Other. Usually, associations with performance ratings.
– Many studies of the impact of economic environment on subjective well-being at work (reviewed in Warr (2007) – Discriminating patterns in the determinants of well-being indices: • e.g. effort is more negatively associated with the contentmentanxiety index than with the enthusiasm-contentment index (Green, 2008) • potential for a more nuanced view of the impact of job characteristics on well-being
– But: no/little evidence on the mobility effects of well-being 5
Does well-being predict mobility better than job satisfaction? • Two possibilities: 1. If “better” indicator of welfare, one would expect well-being indicators to predict quitting better than job satisfaction 2. But, job satisfaction can be interpreted as an evaluation of the job against alternatives (LG & M), hence a better indicator or the utility differences between current and alternative jobs, hence more closely correlated with quit propensity. With similar force, some job satisfaction items closer to happiness as self-validation rather than “experience happiness”. 6
Specific research questions: 1. Do well-being indicators predict voluntary mobility, and if so: 2. Among well-being indicators, is the Enthusiasm-Depression or the ContentmentAnxiety axis more closely linked with mobility? 3. Is it, specifically, low well-being that induces mobility? 4. Which type of indicator best predicts job mobility: Job satisfaction or Subjective wellbeing? 7
Data • The 2001 and 2006 UK Skill Surveys (part of a series of X-sections) • Aged 20-60/65; in work; random probability samples; reasonable response rates; nationally representative • 2001 postal follow-up in 2002: 3,262 usable
Table 1. Stay Leave without unemployment Leave with unemployment
82.4% 11.2% 6.4% 8
Well-being measures Thinking of the past few weeks, how much of the time has your job made you feel each of the following...?) Calm Tense Contented Relaxed Uneasy Worried Enthusiastic Cheerful Depressed Gloomy Miserable Optimistic
C-A scale, alpha of 0.81
Response Scale •Never •Occasionally •Some of the time •Much of the time •Most of the time •All of the time
E-D scale, alpha of 0.80
9
.5 .4
.5
.3
.4
.2
Density
.3
.1
.2
0
.1 0
6
1
2
3 4 5 Contentment-Anxiety Scale Mean: 3.73; sd: 0.90; skewness: 0.08
6
10
20
30
40
3 4 5 Enthusiasm-Depression Scale Mean: 4.31; sd: 0.83; skewness: -0.55.
0
2
Percent
1
0
2 4 Overall Job Satisfaction Mean: 4.24; sd: 1.20; skewness: -0.92
6
10
Well-Being in Occupations 5
4.5
4
3.5
3
2.5
2
Enthusiasm-Depression 1.5
Contentment-Anxiety
1
0.5
0
11
an uf ac Co tur W ns ing ho tr Ho lesa u ct io le te n ls & & Re Re Tr ta an st i Re au l s po al ra rt Es nt & s ta St te or & ag Bu e F si i n an Pu nes cia s bl S l ic Ad e rvi ce m s in ist ra He tio al E th du n & c So atio n O cia th l er W o C om rk m un ity
M
Well-Being in Industries
5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Enthusiasm-Depression
Contentment-Anxiety
12
Changes over 1992-2006
Correlation Matrix
E-D
4.6 4.4
J-S
1992
E-D
2001 4.2
C-A
1
2006
C-A 0.603
4 3.8
J-S
1
0.590 0.440
1
3.6 3.4 Enthusiasm-Depression
Contentment-Anxiety
Job Satisfaction
13
Method: multinomial logit analysis Table 2a Well-Being and “Quitting” Marginal Effects on the Predicted Probability of Quitting
Enthusiasm-Depression
(1)
(2)
-0.0203***
-0.0200***
(0.00645)
(0.00618)
Contentment-Anxiety
(3)
(4)
-0.00821
-0.0106*
(0.00620)
(0.00587)
CONTROLS
NO
YES
NO
YES
Observations
3228
3228
3228
3228
Log likelihood
-1860
-1726
-1867
-1730
14
Table 2 Well-Being and Mobility
VARIABLES
Enthusiasm-Depression
Contentment-Anxiety CONTROLS Observations Log likelihood
(1) “Quit”
(2) “Involuntary mobility”
(3) “Quit”
(4) “Involuntary mobility”
-0.0236***
-0.0148**
-0.0207***
-0.00719
(0.00813)
(0.00620)
(0.00782)
(0.00528)
0.00506 (0.00770)
0.00266 (0.00593)
0.00110 -0.00399 (0.00734) (0.00489) YES 3228 -1726
NO 3228 -1859
Marginal Effects on the Predicted Probability of Each Mobility Outcome Multinomial logit estimates. *** p
