The Value of Nascent Skills for Employability in Peru Informing Human Development: ESW Fair Wednesday, January 12, 2011
Dec 31, 2015
The Value of Nascent Skills
for Employability in Peru
Informing Human Development: ESW Fair
Wednesday, January 12, 2011
Motivation and Questions
1. Why focus on multiple skills? (Beside schooling)
2.Which skills matter most for employability in Peru? (New Skills & Labor Survey)
3.How can they be developed through public intervention? (What it means for the Bank)
3
Why multiple skills? Peruvian employers want both cognitive and socio-emotional skills
• Corroborated by data from ICA surveys, public job intermediation service, qualitative study/interviews of larger firms (similar to OECD, other MICs)
~=40%Non-cognitive
Wage costs
Lack trained staff
Dishonesty/unreliability
Ineptitude
Irresponsibility
Workers too slow
Others
0 5 10 15 20 25 30 35
Personal references
Years of experience
Police report
Age
Gender
Secondary education
Technical education
Primary education
Family situation
University education
Religious beliefs
Other
0 2 4 6 8 10 12 14 16 18 20
Employers’ reported problems to hire suitable workers (% responses)
Employers’ reported factors considered always/frequently to assess workers suitability (% responses)
Source: Peru Firm Informality Survey 2007, N=804 firms,1-50 employees
Which skills matter most for employability in Peru?:
Learning from a New Labor Skills Survey
Measuring Skills and Employability• Developed over 1+ year (DECRG grant), interdisciplinary team
• Representative of main urban areas (n=2,666 HHs) and regions. Built on national HH survey, supplemented by modules on:
– Employability outcomes (employment, earnings, job satisfaction)– Labor insertion, educational trajectories, family background – Skills: Cognitive (receptive language, verbal fluency, working memory,
numeracy-problem solving) and socio-emotional (Big-5 Personality Factors, GRIT)
• Big-five: wide consensus that personality traits cluster into five factors:
– Openness to experience; Conscientiousness; Extraversion; Agreeableness; Neuroticism (inverse of emotional stability)
• GRIT: Narrower trait “perseverance (duration of effort) & passion for long-term goals (consistency of interest)” (Duckworth et al 2007)
• Strong predictor of high achievement in US (over cognitive ability)
6
While interrelated, these skills capture distinct dimensions of human capacity/motivation
Cognitive skills more highly interrelated (‘G’ IQ), socio-emotional less so
• OECD evidence of causal connection behind correlations (Heckman
et al)
Correlation between socio-emotional test scores
PPVTWorking memory
Numeracy-problem solving
Verbal fluency
Summary Cognitive
score
Years schooling
Working memory
0.35 1
Numeracy-problem solving
0.55 0.38 1
Verbal fluency
0.43 0.31 0.43 1
Summary Cognitive score 0.80 0.66 0.81 0.72 1Years schooling 0.62 0.42 0.51 0.40 0.64 1
0.00
0.10
0.20
0.30
0.40
0.50
Open
ness
Extra
versi
on
Agree
ablen
ess
-co
opera
tion
Agree
ablen
ess
-kin
dnes
s
Emoti
onal
stabil
ity
Cons
cienti
ousn
ess
Cogn
itive a
bility
Years
scho
oling
GRIT total Cognitive ability Years schooling
Correlation between cognitive test scores
Which skills matter most for employability in Peru?:
Earnings returns to skills and schooling
Cognitive and socio-emotional skills correlate significantly with earnings
• When assessed individually without controlling for schooling (yes, parental education, demographics), workers scoring 1 std dev higher in these skills earn more:
– 10 percent (working memory, verbal fluency) to 18 percent (receptive language, numeracy)
– 8 percent (Big-five emotional stability, openness to experience, extraversion) to 13 percent (Grit-perseverance )
• Those scoring higher in agreeableness (facet cooperation) have 5 percent lower earnings (found in U.S too!)
These correlations reflect direct earnings returns besides years of schooling
VARIABLES Coefficients from 2-step IV Mincer regression
Years of schooling 0.069***
Overall cog ability 0.094***
Extraversion 0.056
Agreeableness-kindness -0.044
Agreeableness-cooperation -0.080***
Conscientiousness -0.027
Emotional Stability 0.049**
Openness -0.002
GRIT consistency of interest -0.002
GRIT perseverance of effort 0.090*
Dep. variable: Log (hourly earnings)Coefficients from two-steps Mincer regression. 1st step regresses cognitive ability and instruments education to compute cognitive ability residualized from education, which is then used in 2nd step (Hansen, Heckman & Mullen 2004). Control variables: work exp. and square, gender, ethnic group, zone of residence, parental background. Personal traits in z-scores. t-statistics in brackets. *** p<0.01, ** p<0.05, * p<0.1. N=1,142. R-squared= 0.20
Cognitive and socio-emotional skills give comparable advantage in life-time earnings, though below college credentials
+ 2SD
+ 2SD
+ 2SD
+ 2SD
Prim. Comp.
+ 1SD
+ 1SD
+ 1SD+ 1SD
Sec. Incomp.
MEAN MEAN MEAN MEAN Sec. Comp.
- 1SD
- 1SD
- 1SD- 1SD
Higher Tech.
- 2SD
- 2SD
- 2SD- 2SD
Higher Univ.
100
130
160
190
220
250
Cognit. ability Agreeable-Coop. Emot. Stab. GRIT-Perseverance Education
S/. t
hous
ands
Women
+ 2SD
+ 2SD
+ 2SD
+ 2SD
Prim. Comp.
+ 1SD
+ 1SD
+ 1SD+ 1SD
Sec. Incomp.
MEAN MEAN MEAN MEAN Sec. Comp.
- 1SD
- 1SD
- 1SD- 1SD
Higher Tech.
- 2SD
- 2SD
- 2SD- 2SD
Higher Univ.
100
130
160
190
220
250
Cognit. ability Agreeable-Coop. Emot. Stab. GRIT-Perseverance Education
S/.
thou
sand
s
Men• Advantage from higher cognitive ability comparable to some socio-emotional skills
• Both can compensate for low schooling
• Education produces largest earnings inequality – college educated have biggest advantage
• Note: Simulations of age-earnings profiles over work life (graduation-65 yrs retirement) for typical workers using Mincer regression parameters (discount rate 5%)
Which skills matter most for employability in Peru?:
skills schooling
Skills beget skills: Cognitive ability strong predictor of educational achievement
• Holds controlling for host of confounding factors. OECD evidence of two-way causal connection
•
Distribution of Summary cognitive scores by education level
Skills beget skills: Socio-emotional skills also predict educational achievement, though less so
• Holds controlling for host of confounding factors. OECD evidence of two-way causal connection
Distribution of GRIT scores by education level
Cognitive and socio-emotional skills appear more binding for college access than financial constraints
Note: Simulations from bivariate probit regressions: Eq1: 1= pursued tertiary education, 0=otherwise; Eq2: 1= enrolled in college, 0= enrolled in technical/non-university. Controls for individual and family factors such as gender, ethnic group, parental/family background, reported SES and scholastic performance during secondary schooling and. Wald test of indep. Eqns: Prob > chi2 =
0.0063
(*) Increasing monetary resources implies changing self-reported socioeconomic status at the time of secondary schooling from low to medium class; increasing abilities implies moving from the bottom to the upper third of the ability distribution.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
Enrollment in higher education University | institute enrollment
Monetary resources Cog. ability GRIT
Change in probability of tertiary education enrollment
Recapping the Evidence
• Schooling (content + credentials) and cognitive and socio-emotional skills are all very valued in the Peruvian labor market
• Significant gaps in these skills between working-age of better-off and worse-off families
• Timely development of cognitive and socio-emotional skills and improved educational achievement go together, and are essential to a more competitive and equitable Peru
How can these skills be developed through public intervention?
What does it mean for the Bank?
Science and Policy Evaluation gives ample room for Cost-effective Public Intervention
• “Nature” vs “Nurture” separation obsolete: Heritability + family influences interact, both matter• Different sensitive periods: Socio-emotional skills more malleable
through adolescence/early 20s
• With adequate support, good parenting and schools can develop cognitive and socio-emotional skills (Durlak et al;
Heckhman & Cunha; Sankoff et al; WB ECD studies):
– “Tools of the Mind” improve pre-school children’s self-control. Universal school-based interventions (K - high-school), youth mentoring (Big Brother/Sister) & training improve Big-five-related skills
• Early investments to compensate initial disadvantage can be costly, but also yield high returns
What does it mean for the Bank?• Help redefine what it takes to be a well-educated person in
the 21st Century• Cognitive and socio-emotional skills determine a person’s “readiness
to learn” through the life cycle
– Numeracy, literacy and academic qualifications a core but not the only output of education systems. Curricula, learning standards and pedagogic practice should also take socio-emotional skills seriously
• Expand policy research (already happening) to underscore evidence base of links between early HD investments (maternal, child nutrition & health, ECD) & employability skills
– Issue: intangibility and long maturity of investments vis-à-vis short political horizons – better outreach, broader social consensus building (Peru video)
• Learn from and build capacity to adapt successful interventions (expertise on developmental, education psychology)
Thank you!