P AY - FOR -S UCCESS F INANCE FOR E ARLY C HILD C ARE AND E DUCATION 1 Performance-Contingent Finance for At-Risk Children in Alexandria Virginia: A Kauffman-PAES.

PAY-FOR-SUCCESS FINANCE FOR PAY-FOR-SUCCESS FINANCE FOR EARLY CHILD CARE AND EDUCATIONEARLY CHILD CARE AND EDUCATION

1

Performance-Contingent Finance for At-Risk Children in Alexandria Virginia:

A Kauffman-PAES Progress Report to the Invest in Kids Working Group

Partnership for America’s Economic Success

Robert H. Dugger

Managing Partner, Hanover Investment GroupAdvisory Board Chairman, Partnership for America’s Economic Success

October 25, 2011Washington DC

Table of Contents

I. Purpose of the Kauffman PAES Early Childhood Finance Innovation Working Group Page 3

II. Overview of “Pay for Success” Social Impact Bond Financing

Page 5

III. Description of PKSE Bonds and an Alexandria VA Smart Beginnings Capital Partnership Page 12

IV. Presentation of PKSE performance based on results from Perry Preschool, Louisiana LA-4, Pennsylvania PKC, and Granite School District Utah Page 22

2

Part I

Purpose of the Kauffman PAES Early Childhood Finance Innovation

Working Group

3

The goal of the Kauffman-PAES Early Childhood Finance Working Group is to develop ways to pay for quality early care and education that create assets which can be traded among investors worldwide and aggregated in asset-backed securities

The possibility of “invest-in-kid bonds” has been discussed for many years To develop marketable “social impact” assets, we need to:

1. Establish strong statistical yield and repayment histories2. Develop workable instrument contracts and institutional frameworks3. Conduct intermediate “proof of concept” financings, which reveal the

strengths and weaknesses of possible approaches 4. Carryout several full Social Impact Bond invest-in-kids financings

To create markets for such instruments, we need to:

1. Work with major financial institution investment banking, wealth management, and philanthropic divisions

2. Attract broad investor and philanthropic interest in the assets3. Obtain state and federal statutory recognition of the assets4. Steadily accumulate asset performance information5. Carryout several successful asset underwritings

4

Part II

Overview of “Pay for Success” Social Impact Bond Financing

5

Motivations for using performance-contingent or “Pay for Success” funding*

Improving performance and lowering costs.o Payment is based on achieving outcome targets.o Focuses government agencies and social service providers on

achieving program objectives and improving performance in a way that is transparent to taxpayers.

Accelerating adoption of new solutions.o Government pays only if program delivers on its promised impact.o Shifts risk of failure (and of wasting taxpayer dollars on programs that

don’t work) to private sector.o Can also break down the budget silos that hinder investment in

prevention. More rapid learning about what works and what doesn’t.

o Ongoing measurement of a program’s impact is a fundamental component of this approach.

*See Jeffrey Liebman, Social Impact Bonds: A Promising New Financing Model to Accelerate Social Innovation and Improve Government Performance, Center for American Progress, February 2011.http://www.americanprogress.org/issues/2011/02/social_impact_bonds.html.

6

Criteria for successful application of this approach*

1. A potential for high net benefits.

2. Measurable outcomes.

3. A well-defined treatment population.

4. A reliable comparison group or counterfactual.

5. Safeguards against harming the treatment population.

7

*See Jeffrey Liebman, Social Impact Bonds: A Promising New Financing Model to Accelerate Social Innovation and Improve Government Performance, Center for American Progress, February 2011.http://www.americanprogress.org/issues/2011/02/social_impact_bonds.html.

Illustrative potential areas of application*

A. Corrections, including juvenile justice.

B. Services for at-risk youth such as those in the foster care system.

C. Homelessness prevention.

D. Employment services.

E. Preventive health care interventions.

F. Kindergarten readiness/third grade reading levels.

8

*See Jeffrey Liebman, Social Impact Bonds: A Promising New Financing Model to Accelerate Social Innovation and Improve Government Performance, Center for American Progress, February 2011.http://www.americanprogress.org/issues/2011/02/social_impact_bonds.html.

Standard Social Impact Bond (SIB) model* includes --

Government Bond-Issuing Organization

PrivateInvestors

Service Providers

3. Performance-based payments

1. Workingcapital

4. Repayment and ROI fromperformance-based

payments

2. Funding foroperating costs

Payment of return on invested capital to investors Repayment of invested capital Government savings must cover full cost of services

9

See Jeffrey Liebman, Social Impact Bonds: A Promising New Financing Model to Accelerate Social Innovation and Improve Government Performance, Center for American Progress, February 2011.http://www.americanprogress.org/issues/2011/02/social_impact_bonds.html.

Standard SIBs, which provide a return on investment and capital repayment to investors, face implementation challenges

1. Unclear returns on an investment intervention

2. Long years between the intervention and the return

3. Inability to link government cost reduction or revenue gain solely to the investment intervention

4. Involvement of multiple and irreconcilable government jurisdictions

5. Inability to obtain political agreement to give up cost savings or allocate cost savings to investors

6. Presence of incentive inconsistencies among parties to the financing

7. Absence of evaluation and administration experience or capacity among investor entities

10

It may be possible to address the implementation challenges facing standard

SIBs by taking intermediate steps to demonstrate the ability of pay-for-success

financing approaches

11

Part III

Description of PKSE Bonds: “Spend for Pre-K to Reduce Special-Ed Cost”

Alexandria VA Smart Beginnings Capital Partnership

12

The Alexandria Smart Beginnings Capital Partnership is an intermediate step in demonstrating the effectiveness of pay-for-success finance

An intervention to improve school readiness and monetize a net benefit which has been heavily researched

Clear linkages between investor intervention and government cost savings

A short time period between investment and measurable return A means to integrate budget benefits of multiple jurisdictions with

political acceptance A framework for contractual commitments among investment

entities and governments A means to unify business and philanthropic support for the

intervention All government cost savings are re-invested back into the

intervention that generated the savings The savings accumulate steadily, increasing the resources invested in

the intervention and clarifying that pay-for-success financing works

13

The Alexandria Smart Beginnings Capital Partnership sets up a PKSE program to pay for quality prekindergarten to increase school readiness and reduce public school special education costs

Alexandria Smart Beginnings Capital Partnership (ASBCP) proposes to provide parent mentors and “pre-k scholarships” to highly at-risk Alexandria 3 and 4 year-olds to reduce Alexandria public school special-education costs

This project creates PKSE (“peek see”) asset that pay for quality-rated pre-k (PK) to reduce special education (SE) costs.

The goal of this project is to examine institutional, governance and contract details necessary for future invest-in-kid asset development

ASBCP “investor-philanthropists” will not receive a cash yield or repayment of principle

14

A. Investor capital contributions

1. Investors contribute to the SBCP2. SBCP funds are used to fund PKSE scholarships and

administration and parent-mentoring costs

B. Special Education cost savings

1. If the SBCP PKSE program works, there will be fewer special education assignments among PKSE scholars and Alexandria special education costs will decline.

2. The reduction in special education costs due to the PKSE program are allocated to fund more PKSE scholarships.

C. Virginia Preschool Initiative (VPI)

1. The VPI program provides matching funds for local area programs on a per child basis.

2. VPI funding will match the amount of PKSE scholarships awarded each year.

15

The Alexandria Smart Beginnings Capital Partnership PKSE program utilizes three funding sources

A. Grants from non-profit philanthropic institutions

1. Charitable institutions can make grants directly to SBCP 2. Grants pay no interest or return of capital to the grantor

B. Project related investments (PRIs) from non-profit entities

1. Charitable institutions can make PRI grants, loans, and bond and stock purchases to invest in the SBCP

2. Interest and dividends can be paid and capital returned3. PRI investments have to satisfy Section 4944 of the Tax Reform Act

of 1969 --a. Primary purpose of PRI investments must be to accomplish charitable purposes b. No significant purpose of PRI investments is income or the appreciation of

property c. No purpose of PRI investments is to accomplish any political purpose forbidden to

private foundations. http://primakers.net/about/primer

C. Investments from private for-profit entities

1. Investors can make loans and buy the bonds and stock of the SBCP2. Interest and dividends can be paid and capital returned to investors

16

The Alexandria Smart Beginnings Capital Partnership PKSE program can accommodate investments from several kinds of investors

A. Scholarships1. From all economically at-risk children (as defined by the Virginia Preschool

Initiative) in Alexandria City, potential PKSE scholars are chosen at random.2. If parents of the selected children agree to participate in the PKSE program, the

children become PKSE scholars.3. Media coverage of the program will strengthen area awareness of early learning

importance.

B. Mentoring1. Mentors are provided to assist parents become better at caring for and teaching

their children. 2. The mentors help the parents choose the pre-k provider that is best for their

children.3. Mentor guidance is shared among parents of young children and has positive

external effects on other parents and families.

C. Parent choice 1. Parents can choose only from quality-rated providers.2. Parent choice strengthens parent involvement in the care and education of their

children. 3. Choice incentivizes pre-k providers to obtain quality ratings, which strengthens the

Alexandria pre-k system.

17

The Alexandria Smart Beginnings Capital Partnership PKSE scholarships are based on the Minnesota Early Learning Foundation model

Proposed Smart Beginnings Capital Partnership PKSE structure (a)

1. Alexandria Investor Philanthropists -- A group of city business leaders and philanthropists agree to establish a capital partnership to fund a PKSE performance-contingent financing approach to increase future workforce strength and improve public school efficiency.

2. Smart Beginnings Capital Partnership (SBCP) -- A 501c3 non-profit corporation. Contributors to this entity are termed “partners” and are contractually subject to annual capital contribution calls to fund next year’s PKSE allocations if previous allocations are achieving contractually agreed-upon reductions in Alexandria public school special education costs.

3. SBCP Fund – A donor-advised fund within the Alexandria Community Trust. The SBCP Fund provides the SBCP share of PKSE scholarship funding. The fund is managed by the board of the SBCP.

4. SBCP Administration and Evaluation – On behalf of the SBCP board and partners, the SBCP director oversees all aspects of PKSE scholar selection, mentoring, pre-k provision, spec-ed assignment and savings calculations, compliance with SBCP contracts and evaluates program performance.

5. Alexandria School System – Randomly selects children for the PKSE program, provides mentors, and administers the PKSE program.

18

Proposed Smart Beginnings Capital Partnership PKSE structure (b)

6. Virginia Prekindergarten Initiative – VPI matches PKSE scholarship amounts. VPI program provides 50% matching funds for quality-rated pre-k for at-risk four year olds through the school system to pre-k providers based on established risk factors.

7. Scholarships and Mentors – Following the Minnesota Early Learning Foundation model, at-risk children are selected randomly to receive a SBCP PKSE scholarships . If the parents agree to the program, mentors are provided and parents select the pre-k provider.

8. Pre-k provider -- Must be a Virginia Star Quality Initiative (VSQI) rated institution.

9. PKSE graduates – As PKSE graduates move into public kindergarten, lower special education assignments and costs are expected. If lower spec-ed assignments and costs are confirmed, SBCP will issue capital calls to SBCP partners to fund the next round of PKSE scholarship grants.

10.Reinvestment of special-ed cost savings – Using the statistical evidence available and the terms of the contract between them, the SBCP board and the Alexandria School System will agree on the amount of special-ed cost savings that will be re-invested in PKSE scholarship for the next year.

19

SBCP Fund

(Donor-advised Fund in ACT)

Alexandria School Special Ed Cost

Savings Are Reinvested in More

PKSE Scholarships for VSQI Pre-K

Virginia Prekindergarten Initiative (VPI)

Funds to Match PKSE Scholarships

For 4-yr olds

VSQI Pre-K generates

Special Ed Savings for Alexandria School

System

PKSE Scholarships

for at-risk Kids and Mentors for Parents

Parents Choose Any VSQI Rated Pre-K Provider

Smart Beginnings

Capital Partnership

Alexandria Investor-

Philanthropists

Alexandria School System PKSE Account

1

2

3

45

6

7

8

9

10

SBCP Administration, Mentoring and

Evaluation

20

Proposed Smart Beginnings Capital Partnership PKSE money flow

Accurate projections of SBCP PKSE performance will require careful analysis of the unique early childhood conditions in Alexandria City.

In the meantime, a sense of the possibilities can be obtained using estimates of the effects of quality preschool on special education costs drawn from studies done in other states.

21

Part IV

Presentation of PKSE Bond Performance Based on Results from Perry

Preschool, Louisiana LA-4, Pennsylvania PKC, and Utah Granite

School District Studies

22

Pre-K Spec-Ed Cost Effect Fundamentals

Early education cost/benefit estimates are obtained in three general ways –

1.Randomized Control Trial In RCTs subjects are randomly assigned to two groups. One group receives the treatment. The other, the “control” group, does not. After the trial, the result characteristics of the treated and untreated subjects are compared and analyzed.

2.Case Control Trial In CCTs, random assignment is not used. Subjects are identified who have very similar characteristics but who fall naturally in the treatment “case” group or the non-treatment “control” group. After the trial, the results characteristics of the two groups are compared and analyzed. CCTs are used in medical research when true experiments with random assignment are impractical or unethical.

3.Statistical Projection Trial In SPTs, mental development tests or other income, health, or demographic characteristics, which reliably predict a result, are used to project the likely result outcomes for a group of subjects. After the trial, the predicted results and actual results of the treatment are compared and analyzed.

23

Pre-K Spec-Ed Cost Effect Fundamentals

Randomized Control Trial (Classic studies) In RCTs, subjects are randomly allocated to receive one or other of the alternative treatments under study. One group receives the treatment. The other, the “control” group, does not. After the trial all the subjects are monitored in the same way. In early childhood research, the two most famous RCTs are the Perry Preschool and Abecedarian studies. In both studies, quality prek reduced special education assignments.

24

Pre-K Spec-Ed Cost Effect Fundamentals

Case Control Trial (Chicago and Louisiana) The most well-known study comparing “case” and “control” groups is the Chicago Child-Parent Centers. In the Louisiana pre-k program for 4-year olds (LA 4) evaluation, a “case” represents a participant in LA 4, and a “control” represents a child with similar characteristics who was not a participant in LA 4. Researchers found that participation in LA 4 is associated with reduced special education placement for both Free and Reduced Lunch (FRL) eligible children as well as those ineligible for FRL.

25http://www.louisianaschools.net/lde/uploads/11515.pdf

Pre-K Spec-Ed Cost Effect Fundamentals

Statistical Projection Trial (Pennsylvania) Comparing statistically predicted results with actual pre-k graduate results was used to estimate the effects of PA’s Pre-K Counts (PKC) program in reducing special ed costs. 80% of PKC children met critical early school success competencies in the Pennsylvania Early Learning Standards (OCDEL, PAELS, 2005) at transition to kindergarten. The gains of PKC children exceeded the kindergarten transition skills of same-aged peers on the BSSI-3 national norms in spoken language, reading, math, classroom behavior, and daily living skills. The projected PKC special education placement rate was 2.4%, significantly lower than the 18% historical special ed placement rate of PA school districts.

26http://www.heinz.org/UserFiles/Library/SPECS%20for%20PKC%202009%20Final%20Research%20Report%20113009.pdf

Pre-K Spec-Ed Cost Effect Fundamentals

Statistical Projection Trial (Utah) Comparing statistically predicted results with actual pre-k graduate results was used to estimate the effects of Granite School District’s pre-k program in reducing special ed costs. The Peabody Picture Vocabulary Test (PPVT) given to 3 and 4 year-olds is used to predict placement in k-3 special education in many regions. A PPVT score of less than 70 is a strong indication that a child will be placed in special ed. Of 696 Granite School District 3 and 4-year olds given the PPVT test, 238 scored below 70 and were projected to almost certainly be assigned to special ed. However, as a result of attending Granite’s prek program, only 7 were assigned to special ed.

27http://www.partnershipforsuccess.org/uploads/20110713_SustainableFinancingModelPresentationtoPAESNationalBusinessLeaderSummitBostonJuly222011.pdf

The following PKSE financial performance projections do not include –

Declines in number of PKSE scholars in K-12 due to out-location, illness or death

Declines in special education costs due to successful entry of spec-ed students into regular classes

In-locations of k-12 students from other regions with PKSE-like programs

28

29

Twenty-Year PKSE Performance Summary Based on Perry Preschool Results

PROGRAM ASSUMPTIONS       Economic cost and effect assumptions    Initial cost of VSQI pre-k per socio-economically at-risk child per year $10,000  Expected growth rate of cost for a pre-k child per year 3.0%

 Initial cost of Special Education per socio-economically at-risk child per year $15,000

  Expected growth rate of cost for a pre-k child per year 3.0%  Based on 1960s Perry Preschool Randomized Control Trial Results    Expected % of socio-economically at-risk children assigned to Special Ed 38%  Expected % of PKSE children assigned to Special Ed 17%     Program PKSE scholarship assumptions  

 Scholarship contribution per investor per year equals cost of pre-k per year $10,000

  Percent increase in contribution 0.0%     Program PKSE operating cost assumptions    Initial administration costs $60,000  Expected growth rate of administration costs 3.0%  Initial mentoring costs per child per year in PKSE $1,000  Expected growth rate of mentoring costs 1.0%  Initial monitoring costs per child per year in K-12 $100  Expected growth rate of monitoring costs 1.0%     Discount rate 3.0%     Investor contribution assumptions    Initial number of investor-philanthropists 75  Number of years of investor contributions 5

 Average investor contribution per year for scholarships and operating costs

$11,005

     PROGRAM RESULTS       Alexandria and Virginia school systems    Present value of special ed cost savings $24,362,100  Present value of VPI contributions $2,625,528  Present value of spec ed savings as a % of PV VPI contributions 928%     Investor results    Present value of total investor contributions $3,753,748

 Present value of spec ed cost savings as a % of PV investor contributions

649%

 Present value of spec ed savings as a % of PV investor and VPI contributions

382%

30

Twenty-Year PKSE Performance Summary Based on Louisiana LA 4 Results

PROGRAM ASSUMPTIONS       Economic cost and effect assumptions    Initial cost of VSQI pre-k per socio-economically at-risk child per year $10,000  Expected growth rate of cost for a pre-k child per year 3.0%

 Initial cost of Special Education per socio-economically at-risk child per year $15,000

  Expected growth rate of cost for a pre-k child per year 3.0%  Based on 2001-2006 Louisiana LA 4 Case Control Trial Results    Expected % of socio-economically at-risk children assigned to Special Ed 14.5%  Expected % of PKSE children assigned to Special Ed 8%     Program PKSE scholarship assumptions  

 Scholarship contribution per investor per year equals cost of pre-k per year $10,000

  Percent increase in contribution 0.0%     Program PKSE operating cost assumptions    Initial administration costs $60,000  Expected growth rate of administration costs 3.0%  Initial mentoring costs per child per year in PKSE $1,000  Expected growth rate of mentoring costs 1.0%  Initial monitoring costs per child per year in K-12 $100  Expected growth rate of monitoring costs 1.0%     Discount rate 3.0%     Investor contribution assumptions    Initial number of investor-philanthropists 75  Number of years of investor contributions 5

 Average investor contribution per year for scholarships and operating costs

$10,928

     PROGRAM RESULTS       Alexandria and Virginia school systems    Present value of special ed cost savings $3,431,025  Present value of VPI contributions $849,325  Present value of spec ed savings as a % of PV VPI contributions 404%     Investor results    Present value of total investor contributions $3,728,472

 Present value of spec ed cost savings as a % of PV investor contributions

92%

 Present value of spec ed savings as a % of PV investor and VPI contributions

75%

31

Twenty-Year PKSE Performance Summary Based on Pennsylvania Projections

PROGRAM ASSUMPTIONS       Economic cost and effect assumptions    Initial cost of VSQI pre-k per socio-economically at-risk child per year $10,000  Expected growth rate of cost for a pre-k child per year 3.0%  Initial cost of Special Education per socio-economically at-risk child per year $15,000  Expected growth rate of cost for a pre-k child per year 3.0%

 Based on 2008 Heinz Foundation Pennsylvania PKC Statistical Projection Analysis  

  Expected % of socio-economically at-risk children assigned to Special Ed 18%  Expected % of PKSE children assigned to Special Ed 2.4%     Program PKSE scholarship assumptions    Scholarship contribution per investor per year equals cost of pre-k per year $10,000  Percent increase in contribution 0.0%     Program PKSE operating cost assumptions    Initial administration costs $60,000  Expected growth rate of administration costs 3.0%  Initial mentoring costs per child per year in PKSE $1,000  Expected growth rate of mentoring costs 1.0%  Initial monitoring costs per child per year in K-12 $100  Expected growth rate of monitoring costs 1.0%     Discount rate 3.0%     Investor contribution assumptions    Initial number of investor-philanthropists 75  Number of years of investor contributions 5  Average investor contribution per year for scholarships and operating costs $10,977     PROGRAM RESULTS       Alexandria and Virginia school systems    Present value of special ed cost savings $13,707,720  Present value of VPI contributions $1,766,038  Present value of spec ed savings as a % of PV VPI contributions 776%     Investor results    Present value of total investor contributions $3,744,718  Present value of spec ed cost savings as a % of PV investor contributions 366%  Present value of spec ed savings as a % of PV investor and VPI contributions 249%

32

Twenty-Year PKSE Performance Summary Based on Granite School District Projections

PROGRAM ASSUMPTIONS       Economic cost and effect assumptions    Initial cost of VSQI pre-k per socio-economically at-risk child per year $10,000  Expected growth rate of cost for a pre-k child per year 3.0%  Initial cost of Special Education per socio-economically at-risk child per year $15,000  Expected growth rate of cost for a pre-k child per year 3.0%  Based on 2011 Granite County Utah Pre-K Statistical Projection Analysis    Expected % of socio-economically at-risk children assigned to Special Ed 20%  Expected % of PKSE children assigned to Special Ed 1.0%     Program PKSE scholarship assumptions    Scholarship contribution per investor per year equals cost of pre-k per year $10,000  Percent increase in contribution 0.0%     Program PKSE operating cost assumptions    Initial administration costs $60,000  Expected growth rate of administration costs 3.0%  Initial mentoring costs per child per year in PKSE $1,000  Expected growth rate of mentoring costs 1.0%  Initial monitoring costs per child per year in K-12 $100  Expected growth rate of monitoring costs 1.0%     Discount rate 3.0%     Investor contribution assumptions    Initial number of investor-philanthropists 75  Number of years of investor contributions 5  Average investor contribution per year for scholarships and operating costs $10,997     PROGRAM RESULTS       Alexandria and Virginia school systems    Present value of special ed cost savings $19,964,250  Present value of VPI contributions $2,279,773  Present value of spec ed savings as a % of PV VPI contributions 876%     Investor results    Present value of total investor contributions $3,751,037  Present value of spec ed cost savings as a % of PV investor contributions 532%  Present value of spec ed savings as a % of PV investor and VPI contributions 331%

33

Total investor contributions

Total investor contributions

35

36

37