Does investing in Does investing in the children of the the children of the poor have a payoff poor have a payoff

to society?to society?Henry M. LevinHenry M. Levin

Georgia State UniversityGeorgia State University

19 February 200919 February 2009

Educational JusticeEducational Justice

Equity in Education is a Moral ImperativeEquity in Education is a Moral Imperative Largely a matter of fairness or justiceLargely a matter of fairness or justice But inadequate education also exacts toll But inadequate education also exacts toll

on society in terms of lost productivity on society in terms of lost productivity and tax revenues and higher costs of and tax revenues and higher costs of public servicepublic service

Goal is to look at educational equity and Goal is to look at educational equity and adequacy as a social investment in terms adequacy as a social investment in terms of costs and benefits.of costs and benefits.

1972 U.S. Senate Report-1972 U.S. Senate Report-Cost of Inadequate Cost of Inadequate

EducationEducationAnalysis for 25-34 Year Old Males, 1970Analysis for 25-34 Year Old Males, 1970 $237 billion lost in lifetime earnings for $237 billion lost in lifetime earnings for

failure to graduate from high school failure to graduate from high school ($1.2 trillion in 2004 dollars).($1.2 trillion in 2004 dollars).

$71 billion in tax revenues lost ($350 $71 billion in tax revenues lost ($350 billion in 2004 dollars).billion in 2004 dollars).

$40 billion in costs to achieve 100 $40 billion in costs to achieve 100 percent graduation ($200 billion in percent graduation ($200 billion in 2004 dollars)2004 dollars)

Benefit-to-Cost Ratio of Almost 2:1.Benefit-to-Cost Ratio of Almost 2:1.

Problems of 1972 StudyProblems of 1972 Study No Reliable Evaluations of Dropout No Reliable Evaluations of Dropout

Interventions.Interventions. No reliable cost data.No reliable cost data. AssumedAssumed 50 percent increase in spending K-12 50 percent increase in spending K-12

would do job.would do job. AssumedAssumed upward ability bias of 25 percent. upward ability bias of 25 percent. Lack of good data sets on education and public Lack of good data sets on education and public

health costs, criminal justice costs, public health costs, criminal justice costs, public assistance costs.assistance costs.

What data did exist did not include covariates What data did exist did not include covariates to adjust for non-educational factors.to adjust for non-educational factors.

ReduxRedux

Revisited Beginning 2004.Revisited Beginning 2004. Research Team Colleagues:Research Team Colleagues:

Clive Belfield, Economics, Queens Clive Belfield, Economics, Queens College, CUNY.College, CUNY.Cecilia Rouse, Economics, Princeton.Cecilia Rouse, Economics, Princeton.Peter Muennig, Public Health, Peter Muennig, Public Health, Columbia.Columbia.

Series of Studies.Series of Studies.

High school dropoutsHigh school dropouts

Many ways to count dropouts but end result Many ways to count dropouts but end result is the sameis the same

Approx. 3 of 10 students are dropoutsApprox. 3 of 10 students are dropouts Rate higher for males than femalesRate higher for males than females For minorities, 4 of 10 are dropoutsFor minorities, 4 of 10 are dropouts US lags most industrialized countries in US lags most industrialized countries in

graduation ratesgraduation rates Dropouts rising, not falling (Heckman 2008)Dropouts rising, not falling (Heckman 2008) Single cohort 20 year olds, 700,000 dropoutsSingle cohort 20 year olds, 700,000 dropouts

How to reduce the dropout How to reduce the dropout raterate

Many factors influence dropoutsMany factors influence dropouts Inadequate educational investment is oneInadequate educational investment is one Search for interventions that have been Search for interventions that have been

demonstrated, using a strong research demonstrated, using a strong research method, to reduce the dropout ratemethod, to reduce the dropout rate

Over 200 references, but few with strong Over 200 references, but few with strong evaluations and resultsevaluations and results

Effective InterventionsEffective Interventions

Longitudinal to link interventions Longitudinal to link interventions with high school graduationwith high school graduation

Use of experimental or strong, quasi-Use of experimental or strong, quasi-experimental designexperimental design

Evaluation implementation of a high Evaluation implementation of a high quality.quality.

Only 5 of more than 200 intervention Only 5 of more than 200 intervention studies met these criteria.studies met these criteria.

Cost Determination and Cost Determination and Cost-EffectivenessCost-Effectiveness

Few evaluations include costs.Few evaluations include costs. Those evaluations that mention costs Those evaluations that mention costs

provide no information on cost provide no information on cost methods used.methods used.

Evaluators typically have little Evaluators typically have little understanding of how to measure understanding of how to measure costscosts

Cost-MethdologyCost-Methdology

Established consistent method based upon Established consistent method based upon accepted economic criteria (1975)accepted economic criteria (1975)

Expansion and applications in Expansion and applications in Cost-Cost-Effectiveness AnalysisEffectiveness Analysis (1983) and (second (1983) and (second edition, 2001).edition, 2001).

Used for Perry Preschool (Barnett 1985).Used for Perry Preschool (Barnett 1985). Used to compare cost-effectiveness of four Used to compare cost-effectiveness of four

interventions: computer-assisted instruction, interventions: computer-assisted instruction, smaller class size, longer school days, and smaller class size, longer school days, and peer tutoring (Levin, Glass, and Meister peer tutoring (Levin, Glass, and Meister 1987).1987).

Steps Required for Steps Required for CostingCosting

Specify resource ingredients necessary for Specify resource ingredients necessary for intervention.intervention.

Determine from reports, observations, Determine from reports, observations, interviews—interviews—

Rarely is detail found in evaluations of Rarely is detail found in evaluations of interventions. interventions.

Establish market price or shadow price of each Establish market price or shadow price of each ingredient.ingredient.

Determine total cost of intervention.Determine total cost of intervention. Determine cost per participant or set number Determine cost per participant or set number

of participants.of participants.

Application to DropoutsApplication to Dropouts Estimate cost per 100 participants.Estimate cost per 100 participants. Divide this cost by the number of “additional Divide this cost by the number of “additional

graduates” attributed to intervention.graduates” attributed to intervention. Add costs of additional years of schooling for Add costs of additional years of schooling for

additional graduates.additional graduates. Add costs of post-secondary education for Add costs of post-secondary education for

estimated transition to higher education of estimated transition to higher education of portion of additional graduates.portion of additional graduates.

Assumes transition to higher education will Assumes transition to higher education will be lower than average—used bottom be lower than average—used bottom quartile in reading.quartile in reading.

Present ValuePresent Value

Convert to present value at Age 20 Convert to present value at Age 20 of overall investment at 3.5 percent of overall investment at 3.5 percent interest rate for comparison with interest rate for comparison with benefits.benefits.

Present values of costs and benefits Present values of costs and benefits can be compared directly.can be compared directly.

Lottery example--$1 million received Lottery example--$1 million received as $50,000 over 20 years or as lump as $50,000 over 20 years or as lump sum.sum.

The benefits of graduationThe benefits of graduation

1. Private benefits to the individual 1. Private benefits to the individual who graduateswho graduates

2. Fiscal benefits to the taxpayer2. Fiscal benefits to the taxpayer Higher tax revenues because of Higher tax revenues because of

increased earningsincreased earnings Lower government expenditures on Lower government expenditures on

health, crime, welfare, remedial health, crime, welfare, remedial education, public serviceseducation, public services

Fiscal benefits per Fiscal benefits per additional high school additional high school

graduategraduate

1) Identify the “causal impact” of 1) Identify the “causal impact” of education on earnings, health, crime, education on earnings, health, crime, and welfareand welfare

2) Calculate the economic benefit to the 2) Calculate the economic benefit to the taxpayer of each “causal impact” taxpayer of each “causal impact” spread over the lifetimespread over the lifetime

3) Expressed as present value at age 3) Expressed as present value at age 2020

Present Value Age 20Present Value Age 20

Like a Certificate of DepositLike a Certificate of Deposit Benefits and costs occur over timeBenefits and costs occur over time Present value takes account of when Present value takes account of when

they are incurred or received and they are incurred or received and tells us what they are worth at point tells us what they are worth at point in time.in time.

Similar to lump sum payment for Similar to lump sum payment for winning lottery instead of 20 years of winning lottery instead of 20 years of annual paymentsannual payments

Impacts: healthImpacts: health

Education is strongly correlated with good Education is strongly correlated with good health, either directly or because of incomehealth, either directly or because of income

High School graduates live 7 years longer High School graduates live 7 years longer than dropoutsthan dropouts

Lifestyle differences-nutrition, health care, Lifestyle differences-nutrition, health care, less substance abuseless substance abuse

Better knowledge and health decisionsBetter knowledge and health decisions Higher income and better jobs mean greater Higher income and better jobs mean greater

health insurance and private coveragehealth insurance and private coverage

Crime:ImpactCrime:Impact

Since 1987 public spending on incarceration has risen Since 1987 public spending on incarceration has risen by 127 percent and on higher education by 21 percentby 127 percent and on higher education by 21 percent

Already several states spend more on incarceration Already several states spend more on incarceration than higher educationthan higher education

Consistent evidence of education on crimes and Consistent evidence of education on crimes and incarcerationincarceration

About half of all incarcerated are high school dropoutsAbout half of all incarcerated are high school dropouts

Focus only on five major crimes (most crimes are Focus only on five major crimes (most crimes are misdemeanors; exclude fraud and juvenile crime)misdemeanors; exclude fraud and juvenile crime)

Impacts: welfare receiptImpacts: welfare receipt Effects of education are strongest for Effects of education are strongest for

those whose dependence on public those whose dependence on public assistance is most intensive such as assistance is most intensive such as single motherssingle mothers

Focus only on three programs: TANF, Focus only on three programs: TANF, housing assistance and food stampshousing assistance and food stamps

Lifetime benefits per Lifetime benefits per additional high school additional high school

graduategraduate

Cost-benefit ratiosCost-benefit ratios

ConclusionConclusion

Increasing high school graduation Increasing high school graduation increases equity and justice, a moral increases equity and justice, a moral commitmentcommitment

Also a great investment for society where Also a great investment for society where the benefits far exceed coststhe benefits far exceed costs

Each additional graduate confers the Each additional graduate confers the equivalent of a CD worth $127,000 to the equivalent of a CD worth $127,000 to the taxpayer beyond personal benefitstaxpayer beyond personal benefits

Schools must choose programs that are Schools must choose programs that are effective to get these resultseffective to get these results

Net Benefits AccumulateNet Benefits Accumulate

Each cohort of 20 year olds has Each cohort of 20 year olds has about 700,000 high school dropoutsabout 700,000 high school dropouts

If we could reduce that number by If we could reduce that number by half, we would provide a present half, we would provide a present value of $45 billion to societyvalue of $45 billion to society

Each additional year would also add Each additional year would also add that amount so that benefits for ten that amount so that benefits for ten cohorts would be almost a half-cohorts would be almost a half-trillion dollarstrillion dollars

Present WorkPresent Work Increase number of interventions in analysis.Increase number of interventions in analysis. Few that measure HS graduation directly.Few that measure HS graduation directly. Many more that increase test scores and that Many more that increase test scores and that

increase 9increase 9thth grade course taking and passing. grade course taking and passing. Estimate impact of improvements in test scores Estimate impact of improvements in test scores

and/or course taking on increase in probability and/or course taking on increase in probability of graduation.of graduation.

One sigma improvement in combined One sigma improvement in combined reading/math scores at eighth grade increases reading/math scores at eighth grade increases probability of graduation by almost 50 percent. probability of graduation by almost 50 percent. Varies among groups.Varies among groups.

THEPRICEWE PAYEconomic and Social Consequencesof Inadequate EducationClive R. BelfieldHenry M. LevinEDITORS (Washington, DC: Brookings, 2007).

ResourcesResourcesCenter for Cost-Benefit Studies in Center for Cost-Benefit Studies in

Education 20 percent discount on bookEducation 20 percent discount on book((www.cbcse.orgwww.cbcse.org))

Henry M. Levin, Teachers College, Henry M. Levin, Teachers College, Columbia University Columbia University

(HL361@columbia .edu)(HL361@columbia .edu)