ellen k. nyhus empar pons university of agder university valencia non-cognitive traits and the...
TRANSCRIPT
Ellen K. Nyhus Empar PonsUniversity of Agder University Valencia
Non-cognitive traits and the gender wage gap
2
Research Question
Do non-cognitive traits contribute towards explaining the gender wage gap?
- Evaluate which part of the gender wage gap is due to gender differences in characteristics (explained part) and which part is due to gender differences in returns to these characteristics (unexplained part)
- Observe how the explained and the unexplained part change with the
inclusion of the non-cognitive traits in the analysis
- Oaxaca-Blinder decomposition of the wage gap
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
Evidence shows that women receive lower wages than men
3
The Big-5 Personality Traits
Extraversion; Emotional Stability; Agreeableness; Conscientiousness and Openness to Experience (Intellect)
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
Locus of Control
INTRODUCTION AND MOTIVATION
We expect internals to have higher wages than externals since they would be more motivated and they would be rewarded for this
Consideration of Future Consequences (CFC)
We expect a positive relationship between CFC and wages. CFC will be incentive enhancing since it will increase the individual’s valuation of keeping a job and therefore make them avoid behaviour that may result in its termination. Employers reward them for this
We analyse the effects of some non-cognitive traits
4
The Big-5 Personality Traits
Based on previous findings regarding productivity and use of hierarchy negotiation tactics we expect to find:
• Positive relationships with wagesEmotional stability
Conscientiousness
Intellect
Negative relationship with wagesAgreeableness
• Ambiguous relationship with wagesExtraversion
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
5
Gender differences• Gender differences in non-cognitive scores:
Schmitt et al (2008): 55 countries. Women report higher levels of neuroticism, extraversion,
agreeableness and conscientiousness
Semykina & Linz (2007): men are more likely to exhibit an internal locus of control and need
for challenge. Women are more likely to exhibit an external locus of control and need for
affiliation
Lund et al (2007): Women are more likely to use social display/networking tactics for getting
ahead in the labour market. Men are more likely to use Deception/Manipulation and
Industriousness/knowledge tactics
• Gender differences in returns to non-cognitive traits:
Nyhus and Pons (2005): agreeableness is associated with lower wages for women
Mueller and Plug (2006): contrary to Nyhus and Pons, men are those punished for
agreeableness in the labour market
Semykina & Linz (2007): returns to internal locus of control and need for affiliation are higher
for women
We expect to find similar gender differences
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
6
Gender wage gap and the non-cognitive traits
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
• Semykina and Linz (2007) for Russia: variation in the scores of
locus of control and need for affiliation/challenge explains as much as
8% of the gender wage gap
• Mueller and Plug (2006) for US: 3-4% of the gender gap is
explained by differences in the Big-Five personality traits, including
differences in traits and trait returns
Our objective is to evaluate the role of locus of control, time preference and the Big-Five personality traits in explaining the gender wage, using Dutch data
7
Data
• DNB Household Survey 2006
• Panel of 2000 households representative of the Dutch population with respect
to socio-economic variables
• Data collected through the Internet panel of CentERdata
• Questionnaires that may be answered in 30 minutes or less are transmitted
to the households on a weekly basis
• The DNB-HS includes detailed information about respondents’ labour market
details, family situation as well as items designed to tap various
psychological concepts
• Due to 1) panel attrition, 2) missing data due to respondents not filling in all
relevant questionnaires, and 3) exclusion of respondents who do not work
(housewife, retired), the sample size was reduced to 252 women, 370 men
DATA AND MEASUREMENT
Ellen K. Nyhus / Empar Pons
8
Non-Cognitive Measures
• Strathman et al’s (1994), “Consideration of Future Consequences” scale, a measure of the extent to which people consider distant versus immediate consequences of possible behaviours.
• Furnham’s (1986), “Economic Locus of Control” scale (ELOC), a measure of locus of control specific to financial behaviours. 2 dimensions
- Internal scale
- Believing in Chance scale
Computed average score for the 10 items comprising each of the factors
• A 50-item scale drawn from the International Personality Item Pool intended to assess the “Big-Five Personality Traits”, Goldberg (1999)
DATA AND MEASUREMENT
Ellen K. Nyhus / Empar Pons
9
1) Obtain the mean scores for men and women
2) OLS estimates of wage equations by gender
LnW natural log of the hourly wage. Annual gross salary divided by the number of working weeks and then by the number of hours worked each week. Employees with more than one employer excluded
X includes
- Human capital• Highest level of education completed (high, middle and low)• Experience in the labour market • Tenure
- Other Controls• Family structure and other socio-economic information • Workplace characteristics
- Non-cognitive traits
'XWln
METHODOLOGY
Ellen K. Nyhus / Empar Pons
10
3) Gender wage gap decomposition How much of the wage gap is due to differences in characteristics and how much is due to different returns to these characteristics?
)(')'(lnln fmfmfmfm XXXWW
Effects due to differences in characteristics: tthe explained part
Effects due to differences in coefficients: unexplained part
Often considered as reflecting discrimination but variables omission problems
METHODOLOGY
Ellen K. Nyhus / Empar Pons
mfffmmfm XXXWW '''lnln
Observed wage gap
+
11
3) Gender wage gap decomposition
Neumark (1988): the non-discriminatory wage is a vector of rates of return
obtained by estimating earnings functions based on the pooled sample of the
two demographic groups, pool
)(')(')'(lnln fpoolfpoolmmpoolfmfm XXXXWW
Explained part Unexplained part
METHODOLOGY
Ellen K. Nyhus / Empar Pons
12
Men (370 observations)
Women (252 observations)
Mean Std. Dev. Mean Std. Dev.
Hourly wage 19.18 7.89 14.83 5.14
Log hourly wage 2.88 0.38 2.64 0.36
Extraversion 29.86 6.99 30.25 6.98
Emotional 36.14 6.52 33.83 5.87
Agreeable 37.51 5.00 40.68 4.80
Intellect 34.88 5.01 33.85 4.84
Conscientious 35.54 5.79 36.97 5.88
Future 50.37 8.40 50.27 8.62
Chance 20.25 5.21 19.44 5.32
Internal 31.80 5.53 31.43 5.30
Men rate themselves as more emotional stable, with more intellect and a higher tendency to believe in chance than women do, while women rate themselves as more agreeable and conscientious than men do
Main statistics
RESULTS
13
Wage equation including human capital and additional controls
men women
Extraversion 0.003 0.002
(0.308) [0.051] (0.515) [0.039]
Emotional -0.003 0.001
(0.348) [-0.046] (0.811) [0.014]
Agreeable -0.006 -0.010
(0.129) [-0.080] (0.046)* [-0.130]
Intelect 0.014 0.005
(0.000)** [0.179] (0.303) [0.068]
Conscientious -0.004 0.003
(0.175) [-0.063] (0.434) [0.048]
Future 0.002 -0.004
(0.327) [0.047] (0.148) [-0.084]
Chance -0.008 -0.006
(0.017)* [-0.111] (0.158) [-0.086]
Internal 0.003 -0.004
(0.281) [0.051] (0.327) [-0.058]
Adjusted R-squared 0.351 0.292
F test (Prob>F) 6,61 (0.000) 1,11 (0.347)
RESULTS
Significant difference in Believing in Chance
14
RESULTS
)(')(')'(lnln fpoolfpoolmmpoolfmfm XXXXWW
Raw differential
0.2457
100%
Explained partDue endowments
(diff. in characteristics)
0.0916
37.3 %
Unexplained partDue coefficients(diff. in returns)
0.1540
62.7%
Ellen K. Nyhus / Empar Pons
Calculating differences in characteristics and cofficients from the
main statistics and the regresion results just presented, we can
obtain the gender wage gap decomposition
15
11.5% are due to differences in non-cognitive traits (Agreeableness
and Intellect) and less than 0.5% to differences in returns
Gender wage gap decomposition:How much the non-cognitive traits explain
RESULTS
Raw differential 0.2457 100 %
Due endowments Due coefficients
Non-cognitive Traits 0.0283 11.5117%
Non-cognitive Traits 0.0009 0.3825%
Extraversion -0.0007 -0.2690 Extraversion -0.0004 -0.1533
Emotional -0.0018 -0.7410 Emotional -0.0021 -0.8663
Agreeable 0.0212 8.6487% Agreeable 0.0036 1.4730
Intellect 0.0122 4.9790% Intellect -0.0015 -0.5970
Conscientious 0.0017 0.7042 Conscientious -0.0032 -1.3083
Future 0.0000 0.0000 Future 0.0018 0.7169
Chance -0.0046 -1.8687 Chance 0.0001 0.0434
Internal 0.0001 0.0584 Internal 0.0026 1.0742
16
Why men earn more than women?
Regarding the main non-cognitive traits effects,
• Women score higher for agreeableness than men do but this trait is punished….so women earn less than men do
• Men score higher intellect than women do and this trait is rewarded ……so men earn more than women do
RESULTS
Ellen K. Nyhus / Empar Pons
17
The inclusion of non-cognitive traits in the wage model
reduces the unexplained gender wage gap in 12.5 points
RESULTS
Gender wage gap decomposition: The non-cognitive traits role
Model without non-cognitive traits
Model with non-cognitive traits
Gender Wage GapRaw differential 0.2457 100% 0.2457 100 %
Explained PartDue EndowmentsDifferences in Characteristics 0.0608 24.8% 0.0916 37.3%
Unexplained PartDue CoefficientsDifferences in Returns
0.1848 75.2% 0.1540 62.7%
18
• Around 12% of the gender wage gap are due to differences in non-cognitive traits (scores and returns)
• Inclusion of the non-cognitive traits allows for a reduction of the unexplained part of the gender wage gap
• A part of the unexplained term in the traditional human capital analysis may not be due to discrimination but rather caused by the omission of important variables as, for example, the non-cognitive traits
CONCLUSIONS
Ellen K. Nyhus / Empar Pons
Thank you for your attention
Ellen K. Nyhus Empar PonsUniversity of Agder University Valencia
Non-cognitive traits and the gender wage gap
20
Background
• Bowles, Gintis & Osborne (2001) argue that some non-cognitive
traits may be incentive-enhancing in the employer-employee relationship, and the employer will reward them
1) Personal self-efficacy (locus of control)
2) Time preference
3) Utility from work
• Previous studies have found relationships between with the Big-Five personality traits and earnings through
1) Productivity
2) Use of Hierarchy Negotiations tactics
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
21
• Goldsmith, Veum & Darity (2000): Internals are typically more motivated to perform than externals, since they perceive themselves to be able to produce desired outcomes
• Groves (2005); Heckman, Stixrud & Urzua (2006); Semykina &Linz (2007): Internals tend to have more success in the labour market in terms of having higher earnings
LC (Rotter, 1966) reflects a generalised belief an individual has
regarding who controls important events in his/her live. Internals: individuals who believe that the outcomes they experience are consequence of their own behaviour. Externals: individuals who believe that the outcomes they experience are consequence of fate
Locus of Control
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
We expect internals to have higher wages than externals
22
Future Orientation
Consideration of future consequences (CFC) is intended to
describe differences in the extent to which a person is likely to consider distant outcomes when choosing his/her present behaviour
CFC will be incentive enhancing since it will raise the individual’s
valuation of keeping a job in the future and therefore make them avoid behaviour that may result in its termination, Bowles, Gintis & Osborne (2001)
We expect a positive relationship between CFC and wages
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
23
The Big-5 Personality traits
•Extraversion: encompasses preference for human contact and attention and the wish to inspire others
•Emotional stability: degree to which an individual may be described as tense, insecure, anxious, depressed and emotional rather than calm
•Agreeableness: degree to which an individual is co-operative, warm and agreeable versus cold, disagreeable and antagonistic
•Conscientiousness: Degree to which an individual is hardworking and organized and tends to follow rules and keep engagements•Openness to experience: person’s propensity to make independent
decisions and their degree of initiative and control. Intellect: aspects such as perceptive, reflective, being imaginative, curious and broad-minded. The Intellect is not a measure of intelligence (small positive correlations with IQ), John & Srivastava (1999)
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
24
Productivity
• Salgado (1997): Meta-study showed relationship between the big-5 personality traits and productivity. Conscientiousness and emotional stability are positively associated with productivity across occupations
• Employers’ behaviour making significant investments in testing the personality of potential employees and using the results in the screening process
The Big-5 Personality Traits AND
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
25
Hierarchy Negotiations Tactics
The Big-5 Personality Traits AND
INTRODUCTION AND MOTIVATION
• The means by which the individual tries to achieve his or her goals in a social environment • A mediating construct between personality traits and outcomes related to the goal of ”getting ahead”
Ellen K. Nyhus / Empar Pons
26
Hierarchy Negotiations tactics
• Deception/Manipulation: Use deceptive self-promotion, derogate others, boast, aggress, use sex, exclude others, ingratiate self with superiors
• Social Display/Networking: help others, cultivate friendship, attract opposite sex, enhance appearance, social participation, enhance appearance, display positive social characteristics
• Industriousness/Knowledge: display knowledge, work hard, advance professionally, obtain education or knowledge, organize, strategize, assume leadership
Kul-Heku & Buss (1996); Lund et al (2007)
The Big-5 Personality traits AND
INTRODUCTION AND MOTIVATION
Ellen K. Nyhus / Empar Pons
27
INTRODUCTION AND MOTIVATION
IndustriousnessKnowledge
WAGES
+ Extraverted + Emotional stability
+Openness/Intellect - Agreeableness + Conscientiousness
Social DisplayNetworking
+ Extraverted +Openness/Intellect
+/- Agreeableness + Conscientiousness
+ Extraverted - Emotional stability
- Agreeableness - Conscientiousness
Deception Manipulation
-+ NS
Ellen K. Nyhus / Empar Pons
Kul-Heku & Buss (1996); Lund et al (2007)
28
•Age effects: regressing the non-cognitive traits variables against age (as well as the second and third order term of age) and keeping the predicted residuals as measures of the non-cognitive traits.
•Labour/life experience effects: We lack the necessary information to control for the fact that labour specific characteristics may affect the non-cognitive variables
We have to be cautious in any causal interpretation of results
DATA AND MEASUREMENT
Non-Cognitive Measures: stability and endogeneity
Ellen K. Nyhus / Empar Pons