Transcript Document

SCHOOLS SKILLS AND SYNAPSES
James J. Heckman
NBER Working Paper No. 14064
June 2008
JEL No. A12
Elena Broccini
American Society is Becoming Polarized…
One aspect of the US increasing polarization is that:
•
a greater percentage of American children is attending
and graduating college: At the same time a greater
percentage is dropping out of secondary school.
•
This produces a growing underclass, neither working nor
going to school.
•
At the same time there is a recent growth in unskilled
migration to the U.S. that increases the proportion of
unskilled Americans in the workforce.
…the Workforce is Less Skilled and Productive
The increase in high school dropouts rate
+
The recent growth in unskilled migration
These two trends reduce the growth in workforce productivity
and produce more people with low skills in US:
• Annual growth in labour productivity has slowed by 0.17%
to 0.35% per year.
• More than 20% of US workforce has so low rate of literacy
that it cannot understand the instructions on vial of pills.
The Argument
• What forces produce these low levels and adverse trends?
• Are the public schools mainly responsible?
• Can we look to school reform to fix the problem?
• Are higher college tuition costs to blame?
 The answer is “No” to all of these questions.
Contrary to prevailing views:
Tuition costs and Schooling Quality explain only trivial
fractions of the gaps in educational attainment by socio
economic status…
Cognitive and Non-cognitive Abilities Matter
•
A substantial body of research shows that cognitive and
non cognitive abilities are important determinants of
schooling and socioeconomic success.
•
By non cognitive abilities we mean socio-emotional
regulation, time preferences, personality factors, the
ability to work with others..
•
They are both equally predictive of many social
outcomes (i.e. earnings, labor force experience, college
attendance, teenage pregnancy, participation in risky
activities, participation in crime, etc).
Ever Been in Jail by Age 30, by Ability (Males)
Source: Heckman, Stixrud, and Urzua (2006).
This figure plots the probability of a given behavior associated with moving up in
one ability distribution for someone after integrating out the other distribution (i.e.
the red line shows the effect of increasing noncognitive ability after integrating the
cognitive ability).
Probability of Teenage Pregancy (Females)
Source: Heckman, Stixrud, and Urzua (2006).
This figure plots the probability of a given behavior associated with moving up in
one ability distribution for someone after integrating out the other distribution (i.e.
the red line shows the effect of increasing noncognitive ability after integrating the
cognitive ability).
Ability Gaps Opens Up at Early Ages and Persist
•
Gaps in the abilities that play such an important
role in determing diverse adult outcomes open up
at early ages across socio economic groups.
•
School plays a minor role in creating or
perpetuating gaps.
•
Even though American children go to very
different schools, depending on their family
background, test scores are remarkably parallel.
Average Percentile Rank on PIAT-Math Score,
by Income Quartile - Unadjusted
Source: Carneiro and Heckman (2003).
The gaps in achievement at age 12 are mostly present at age 6 when children enter
school. High scores are associated with higher socio-economic status.
Average Percentile Rank on PIAT-Math Score,
by Income Quartile – Adjusted*
Source: Carneiro and Heckman (2003).
* When one controls for early family background factors (mother's education
and ability) using regression analysis, the gaps greatly diminish.
Average Anti-Social Behavior Score Percentile,
by Income Quartile - Unadjusted
Source: Carneiro and Heckman (2003).
In this figure an high score is an indicator of behaviour problems. Again gaps open
up early and persist. High scores (worse behaviour problems) are associated with
lower economic status.
Average Anti-Social Behavior Score Percentile,
by Income Quartile- Adjusted*
Source: Carneiro and Heckman (2003).
Again: *controlling for family environments factors, using regression analysis,
the gaps greatly diminish.
Family Environments matter
•
Such regression adjustment cannot a establish causality a
causal interpretation of this evidence, that is instead
supported by experimental evidence.
•
The regression adjustments establish the importance of
family factors in explaining ability gaps.
•
This is a source of concern becuase a growing fraction of all
American children are born into disadvantaged families.
•
Those born into disadvantaged environments are
receiving relatively less stimulation and child
development resources than those from advantaged
families.
Enriching Early Environments
Can Compensate for Early Adversity
• Recent evidence suggests that early environments
play a powerful role in shaping adult outcomes.
• Even though a lot of children are growing up in
adverse environments and this will have negative
consequences for American Society, the good news
is that environments can be enhanced to promote
the quality of children
• American Society must not passively watch its own
decline but policy can matter.
Experimental Evidence: Perry Program
The Perry Pre-school Program enriched the lives of low
income black children, with initial IQs below 85 at age 3, in
Ypsilanti-Michigan between 1962 and 1967.
The program consists on:
• 2.5 hours classroom session per day on weekday mornings
• A weekly 90 minutes home visits by teacher on weekdays
afternoon
• The length of each preschool year was 30 weeks.
• The groups have beeen followed through age 40.
Experimental Evidence: Perry Program Results
IQ scores become stable by age 10, suggesting a sensitive period for their
formation below age 10. On average, the later remediation is given to a
disadvantaged child, the less effective it is.
Experimental Evidence: Abecedarian Program
The Abecedarian Program shows a similar pattern.
• The Program studied 111 disadvantaged children, born
between 1972 and 1977, whose families scored high on
a risk index.
• The mean age at entry was 4.4 months.
• The program was a year-round, full-day intervention
that continued through age 8.
• The children were followed through age 21, and an age
30 follow-up study is in preparation.
Experimental Evidence: Abecedarian Results
The first years are critical and sensitive periods as in the Perry Program
Observations on these empirical studies
1. Skills beget skills and capabilities foster future capabilities.
2. Early learning confers value on acquired skills, which leads
to self-reinforcing motivation to learn more.
3. Early mastery of a range of cognitive, social, and emotional
competencies makes learning at later ages more efficient and
therefore easier and more likely to continue.
 SELF-PRODUCTIVITY: the capabilities produced at one
stage augment the capabilities attained at later stages.
 DYNAMIC COMPLEMENTARITY: capabilities produced
at one stage of the life cycle raise the productivity of
investment at subsequent stage.
Return to a marginal increase in investment
at different stages of the life cycle
Early investments have the
highest return; due to dynamic
complementarity they must
be
followed
by
later
investment if maximum
value is to be realized.
This figure shows the return to a marginal increase in investment at different
stages of life cycle starting from a position of low but equal initial
investment at all ages.
Later Remediation is Costly
What if we do not invest early?
•
Remedial interventions for disadvantaged adolescents
who do not receive e strong initial foundation of skills
have low rates of return.
•
As currently congured, public job training
adult literacy services, prisoner rehabilitation
and education programs for disadvantaged
current levels of expenditure produce low
returns.
programs,
programs,
adults at
economic
 Remediation is costly, but it is not impossible, except
when we get to very low levels of initial conditions.
Returns to one more euro of investment
as perceived initially and at age 3
Return to an extra dollar of
investment as viewed at age 3 if
optimal investment is made in the
first three years (complementarity
not too strong) and a dollar of
investment is made at all ages (and
is assumed to be less than the
equilibrium amount)
Return to an extra dollar of investment
as viewed at age 3 if suboptimal
investment is made in the first three
years and a dollar of investment is
made at all ages (and is assumed to be
less than the equilibrium amount).
A Model of Investment in Human Capabilities
Cunha and Heckman(2008) and Cunha, Heckman and Schennach (2007) estimate
atechnology of capability production to understand how the skills of the children
evolve in response to:
1. The stock of skills children have already accumulated (θ)
2. The investment made by their parents (I)
3. The stock of skills accumulated by parents (h)
•
When the child is t years old, the stock of capability is made by:
•Substituting for
stage t + 1,
θt, θt-1,…, repeatedly, one can rewrite the stock of capabilities at
θt+1, as a function of all past investments:
A model of Investment in Human Capabilities
Dynamic complementarity arises when:
Self productivity arises when:
•
•
•
The joint effects of self-productivity and dynamic complementarity help to
explain the high productivity of investment in disadvantaged young children but
the lower return to investment in disadvantaged adolescent.
for those disadvantaged children the stock of capabilities is low and hence the
complementarity effect is lower.
Suppose, for analytical simplicity, that there are 2 stages of childhood, (T = 2).
The adult stock outcome is given by:
A model of Investment in Human Capabilities
•
A general technology is the CES (Constant Elasticity of Substitution)
production function:
•
for
and
• 
is a measure of how well late inputs substitute for early inputs. It governs
how easy it is to compensate for low levels of stage 1 investment in producing
later adult capabilities.
•
γ
is a capability multiplier. It captures the productivity of early investment not
only in directly boosting h’ (through self-productivity) but also in raising the
productivity of I2 by increasing θ2 through first period investments. Thus I1
directly increases θ2 which in turn affects the productivity of I2 in forming h’. It
captures the net effect of I1 on h’ through both self-productivity and direct
complementarity.
A Model of Investment in Human Capabilities
Consider 2 polar cases:
When
When
•
I1 and I2 are perfect substitutes
•
I1 and I2 are perfect complements
•
Early deficits can be perfectly
remedied by later intervention.
However it may be not cost effective.
•
•
Early deficits cannot be remedied.
If investment in period 1 are very
low, no remediation is possible.
Adult human capital and
consequently adult succes is defined
in the first period of life.
•
•
A Model of Investment in Human Capabilities
More generally, when  is small:
• Low levels of early investment I1 are not easily remediated
by later investment I2.
• High early investment should be followed with high late
investment if the early investment is to be harvested.
Technology explains why returns to education are low in the
adolescent years for disadvantaged adolescents (low h, low
I1, low θ1).Without the proper foundation for learning (high
levels of θ2) in technology adolescent interventions have low
returns.
Conclusions
• The optimal policy is to invest relatively more in the
early years. But early investment must be followed up
to be effective.
• Later remediation is possible but to attain what is
accomplished by early investment is much more
costly.
• If society intervenes too late and individuals are at too
low a level, later investment can be economically
inefficient.
Practical Issues
• Who should be targeted? Disadvantages children who do
not receive substantial amounts of parental investment in the
early years.
• With what programs? Those targeted the early years.
• Who should provide the programs? Engage private industry
and other social groups that: a) draw in private resources; b)
create community support; c) represent diverse points of
view.
• Who should pay for them? Programmes should be universal
to avoid stigmatization, even though a problem of
compliance can persist.
Not Only in the US...
Young males not in employment, education
or training in the OECD
Young females not in employment, education
or training in the OECD
Source: OECD social indicators, 2009
Source: OECD social indicators, 2009
2002
2006
Variation
Australia
Austria
Belgium
Canada
Czech Republic
Denmark
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Japan
Luxembourg
Netherlands
New Zealand
Norway
Poland
Portugal
Slovak Republic
Spain
Sweden
Switzerland
UK
USA
6,9
6,7
-0,1
8,1
7,1
…
7,3
7,5
0,2
8,7
8,0
-0,7
5,8
4,7
-1,1
2,4
4,8
2,4
..
4,1
…
3,7
6,7
3,0
4,3
4,1
-0,1
5,2
7,9
2,7
8,3
6,4
-1,9
5,2
5,3
0,1
10,8
12,2
1,4
8,3
7,5
-0,8
..
4,4
..
4,0
3,3
-0,7
..
7,7
…
..
3,5
…
3,5
3,8
0,4
7,7
7,8
0,1
17,7
6,5
-11,2
6,9
9,6
2,7
5,9
6,2
0,3
5,8
7,7
1,9
8,2
11,5
3,2
6,4
6,0
-0,4
Australia
Austria
Belgium
Canada
Czech Republic
Denmark
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Japan
Luxembourg
Netherlands
New Zealand
Norway
Poland
Portugal
Slovak Republic
Spain
Sweden
Switzerland
UK
USA
2002
2006
Variation
7,2
7,4
0,2
4,4
6,0
1,6
6,4
6,7
0,3
7,2
6,7
-0,5
6,3
4,3
-2,0
2,4
4,1
1,6
..
3,0
…
3,2
6,2
3,0
5,1
4,3
-0,9
7,5
9,8
2,3
7,8
5,6
-2,2
4,5
4,7
0,2
10,3
11,4
1,1
10,6
10,5
-0,1
..
…
3,7
2,6
-1,1
..
8,9
…
..
3,4
…
2,6
3,7
1,1
6,8
7,7
0,9
13,5
6,8
-6,6
7,5
10,6
3,1
3,3
4,3
1,1
5,8
7,5
1,6
8,9
10,3
1,4
7,5
6,7
-0,8
…but no (or little) emphasis on early years
and skill formation policy
• As in the USA in many OECD countries there is an
increasing young underclass neither working nor going to
school.
• Currently in the OECD conuntries there is substantial
spending on active labor market programs: ALMP.
• A large array of studies surveyed in Heckman, LaLonde, and
Smith (1999) and Martin and Grubb (2001), as well as more
recent studies, show that:
• Few ALMP lift most participants out of poverty.
• They are primarily focused on older workers (older than 25).