Research | Training | Policy | Practice
Reducing the Effect of Poverty through Early Childhood
Interventions Katherine Magnuson
University of Wisconsin-Madison Institute for Research on Poverty and School of
Social Work
Institute for Research on Poverty Webinar February 20, 2013
Talk Outline
• Stylized facts about “disadvantage” and inequality
• Differences in achievement by early disadvantage
• Explanatory pathways & developmental timing
• Discussion of effective policy and programmatic responses to poverty
What is Poverty?
• Having little or no money or means of financial support
• Federal Government created a set of poverty thresholds in the 1960s – In 2012, the poverty threshold for family of three was
$19,090 and for a family of four it was $23,050 – 2011 US poverty rate was 15%
• Alternative poverty measures are increasingly recognized as important
Fifteen-year Poverty Experiences of Children in the Panel Study of Income Dynamics born between 1975-1987, by Race and Maternal
Characteristics at Birth
Ave. Number of Years
Poor
Never Poor Poor for at least 5 years
Poor for at least 8 years
Total Sample 1.81 65% 15% 10%
African American 5.53 30% 46% 37%
White 0.93 75% 7% 4%
Unmarried Mother 5.39 24% 46% 33%
Mother Education: < High School
Degree
5.03 31% 44% 33%
Notes: Calculations of the Panel Survey of Income Dynamics conducted by Kathleen Ziol-Guest. Figures in this table are based on weights that adjust for differential sampling and response rates.
Thursday, March 21, 2013 6
Low-income children enter school with low-levels of academic skills & these differences persist
Low-income children enter school with higher levels of problem behavior & these differences persist
Family income relative to 1970
0.60
0.80
1.00
1.20
1.40
1.60
Inco
me
rela
tive
to 1
970
Top 5%
Top 20%
Bottom 20%
1970= 100
Source: Duncan & Murnane (2011) calculations based on data from the U.S. Bureau of the Census. Shaded areas indicate recession years
Achievement gaps by race and income, by birth year
0.00
0.50
1.00
1.5019
4019
4319
4619
4919
5219
5519
5819
6119
6419
6719
7019
7319
7619
7919
8219
8519
8819
9119
9419
9720
00
Dif
fere
nce
in s
tand
ardi
zed
test
sc
ores
Black/white gap based on NAEP data
Income gap based on various national surveys
Source: Reardon, 2011
Figure 3: College graduation rates for low and high income children
5% 9%
36%
54%
0
25
50
75
1975-78 1993-96
Perc
ent c
ompl
etin
g co
llege
Year turned 14
Low income
High income
Source: Calculations based on Bailey and Dynarski (2011). Low and high incomes are defined as the bottom and top quartiles of the parent income distribution.
How much does childhood poverty really matter?
• From a scholarly perspective it’s hard to know…
• “When we ask about the relationship between poverty and child outcomes it is not completely clear whether we are asking about the low income of poor families or the complex set of circumstances that results in low income.” Mayer (2010)
• Theory: – Increasingly good theoretical models (pieces of
which have been confirmed)
• Empirically: – Never had a good experiment of income support – But, lots of clever attempts to use quasi-
experiments to answer the question • On average these show larger effects than non-clever
approaches • Some puzzles to figure out
How much does childhood poverty really matter?
An Explanatory Model
Child-hood poverty
Stimulation
Cognitive stimulation in the home
Type and quality of child care
Quality of schools and neighborhood
Stress
Maternal mental health
Parenting
Child Achievement, Behavior and
Health
Brain & physiological development
Adult Achievement, Behavior and
Health
Contexts: neighborhoods, schools, peers
Making Work Pay Pays off in Student Achievement
Source: Morris et al. (2005)
Effects on Student Achievement
vs. No Intervention
Earning Supplement
All Other Programs
.100
.075
.050
0 Earning Supplement
All Other Programs
Age 2-3 Age 4-5
.025
Expansions in EITC during 1990s
Source: Dahl & Lochner, 2009
Effects of EITC expansion on children’s test scores and mothers’ health
.20sd* .17sd*
-8.3%* -.08*
-0.20
0.00
0.20
Math Reading
Dahl and Lochner, forthcoming; Evans and Garthwaite, 2009
Children Mothers
CRP ≥ 0.3 mg/Dl
# bad mental
health days
Distribution of Casino Revenues to American Indian Families
Source: Akee et al., 2009
Distribution of Casino Earnings (Akee eta al., 2009)
• Panel of 1420 children (ages 9-13 followed through ages 19-21) – American Indian n=350, non-Am. Indian n=1,070
• Compare AI and non-AI cohorts of children with casino income to those without (difference in difference approach)
Casino Earnings Improve Youth Educational Attainment
Estimated Effect of Having a AI Parent
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
Not Poor Family Poor Family
Year
s of
Com
plet
ed S
choo
ling
Source: Akee et al., 2009
MDRC’s evaluation of CCT (Riccio et al., 2010)
• 2007-2010 in low-income NYC neighborhood • Provide financial incentives for positive
behavior – School attendance, health visits, employment – Average increase in income >3,000 – Reductions in poverty and extreme poverty
• Few main effects on children’s school outcomes – Achievement, attendance
• But focuses on early implementation
Notable gaps…
• Health • Behavior • Persistence of effects over time • Developmental timing
The Importance of Early Childhood
Neural Circuits are Wired in a Bottom-Up Sequence
(700 synapses formed per second in the early years)
Source: C.A. Nelson (2000)
FIRST FIVE YEARS
1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Months Years
Sensory Pathways (Vision, Hearing)
Language Higher Cognitive Function
Early Life Experiences Are Built Into Our Bodies
(For Better or For Worse)
Stimulation: Stable and supportive relationships, language-rich environments, and mutually responsive, “serve and return” interactions with adults promote healthy brain architecture and adaptive regulatory systems.
Stress: Excessive or prolonged activation of stress response systems and reduced availability of the buffering protection of supportive relationships can weaken brain architecture and disrupt the development of other organ systems.
Neural correlates of socioeconomic status in the developing human brain
Developmental Science Volume 15, Issue 4, pages 516-527, 29 MAR 2012 DOI: 10.1111/j.1467-7687.2012.01147.x http://onlinelibrary.wiley.com/doi/10.1111/j.1467-7687.2012.01147.x/full#f1
From Duncan et al., 2012
Programs and Policy Responses • Emphasis on employment conditioned income
supports to improve family income • EITC, Income Supplements, casino revenues
• But, how easily generalized to other types of programs?
• Some might seem like small leaps Child Tax Credit, SNAP
• Others involve other incentives and experiences and thus are likely not direct corollaries (child care subsidies)
• Do these programs increase stimulation & reduce stress?
• How much money? Conditioned on what? Paid how often? To whom?
Programs and Policy Responses
• Early Childhood Education for 3-5 year-olds • Consistent evidence of short-run effects non-
experimental evidence of long-run benefits from several program models
• Complicated comparisons across program models • More attention to learning how to improve
program quality
0
20
40
60
80
Per
cent
atte
ndin
g pr
esch
ool
1970 1980 1990 2000 2010 Year
Income Quartile1 Income Quartile 2
Income Quartile 3 Income Quartile 4
by family income quartile Percent of children 3- and 4-year-olds enrolled in preschool
29 -0.50
0.00
0.50
1.00
1.50
2.00
1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Ave
rage
eff
ect s
ize
in s
d un
its
Average ECE impact at end of treatment
Head Start Non Head StartPerry
Abecedarian
National Head Start
Innovation & Uncertainty
• Home visiting programs • Much to learn about how to promote positive parent-
child interactions
• Approaches in very early childhood (birth-3) – How to balance income support & employment
• Place-based approaches (CHZ) • Combining approaches…
– Where should building parents’ human capital fit in?