Implementation Completion and Results Report (ICR) Document · 2020. 7. 17. · document of the...

Document of The World Bank

FOR OFFICIAL USE ONLY

Report No: ICR00004508

IMPLEMENTATION COMPLETION AND RESULTS REPORT

(Loan 8354-CO)

ON A

LOAN

IN THE AMOUNT OF US$200 MILLION

TO THE

INSTITUTO COLOMBIANO DE CREDITO EDUCATIVO Y ESTUDIOS TECNICOS EN EL EXTERIOR (ICETEX)

WITH THE GUARANTEE OF THE REPUBLIC OF COLOMBIA

FOR AN

ACCESS WITH QUALITY TO HIGHER EDUCATION PROJECT-ACCES II -SOP PHASE 2 (P145782)

August 19, 2019

Education Global Practice

Latin America and Caribbean Region

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

CURRENCY EQUIVALENTS

Exchange Rate Effective August 2, 2018

Currency Unit = Colombian Pesos (COP)

COP 2,900 = US$1

FISCAL YEAR July 1 - June 30

ABBREVIATIONS AND ACRONYMS

ACB Appraisal Cost-Benefit

ACCES Access with Quality to Higher Education (Acceso con Calidad a la Educación Superior)

APL Adaptable Program Loan

APOTEOSYS ICETEX Disbursement Management Software

B-40 Bottom 40 percent of the Income Distribution

CPS Country Partnership Strategy

C&CTEX ICETEX Portfolio Management Software

DANE National Directory of Statistics (Dirección Nacional de Estadística)

DNP National Planning Department (Departamento Nacional de Planeación)

GDP Gross Domestic Product

GoC Government of Colombia

HEI Higher Education Institution

ICL Income-Contingent Loans (Financiamiento contingente al ingreso)

IERR Internal Economic Rate of Return

IRI Intermediate Result Indicator

ICETEX Colombian Institute for Educational Credit and Technical Studies Abroad (Instituto Colombiano de Crédito Educativo y Estudios Técnicos en el Exterior)

ICR Implementation Completion and Results Report

IFC International Finance Corporation

IPP Indigenous Peoples Plan

IRI Intermediate Results Indicator

IRR Internal Rate of Return

IT Information Technology

MEN National Education Ministry (Ministerio de Educación Nacional)

OECD Organisation for Economic Co-operation and Development

PACES Access and Quality in Higher Education Project

PDO Project Development Objective

PDOI Project Development Objective Indicator

PND National Development Plan (Plan Nacional de Desarrollo)

PNDE National Decennial Education Plan (Plan Nacional Decenal de Educación)

SENA National Learning Service (Servicio Nacional de Aprendizaje)

SISBEN Identification System for Social Program Beneficiaries (Sistema de Identificación de Beneficiarios de Programas Sociales)

SNIES National Information System for Higher Education (Sistema Nacional de Información de la Educacion Superior)

SPADIES System for the Prevention of Dropouts in Higher Education (Sistema para la Prevención de la Deserción de la Educacion Superior)

SOP Series of Projects

T&T Professional technical and technological higher education institutions

UBN Unsatisfied sent the quality

UCP Project Coordination Unit (Unidad Coordinadora de Proyecto)

Regional Vice President: Axel van Trotsenburg

Country Director: Ulrich Zachau

Senior Global Practice Director: Jaime Saavedra Chanduvi

Practice Manager: Emanuela Di Gropello

Task Team Leader(s): Pedro Cerdan-Infantes

ICR Main Contributor: Uriel Kejsefman

TABLE OF CONTENTS

DATA SHEET .......................................................................................................................... 1

I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES ........................................................... 5

A. CONTEXT AT APPRAISAL .........................................................................................................5

B. SIGNIFICANT CHANGES DURING IMPLEMENTATION ................................................................9

II. OUTCOME ....................................................................................................................... 10

A. RELEVANCE OF PDOs ............................................................................................................ 10

B. ACHIEVEMENT OF PDOs (EFFICACY) ...................................................................................... 11

C. EFFICIENCY ........................................................................................................................... 16

D. JUSTIFICATION OF OVERALL OUTCOME RATING .................................................................... 17

E. OTHER OUTCOMES AND IMPACTS ......................................................................................... 17

III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME .................................. 19

A. KEY FACTORS DURING PREPARATION ................................................................................... 19

B. KEY FACTORS DURING IMPLEMENTATION ............................................................................. 19

IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME .... 20

A. QUALITY OF MONITORING AND EVALUATION (M&E) ............................................................ 20

B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE ..................................................... 21

C. BANK PERFORMANCE ........................................................................................................... 22

D. RISK TO DEVELOPMENT OUTCOME ....................................................................................... 23

V. LESSONS AND RECOMMENDATIONS ................................................................................ 24

ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS ........................................................... 26

ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION ......................... 43

ANNEX 3. PROJECT COST BY COMPONENT ........................................................................... 45

ANNEX 4. EFFICIENCY ANALYSIS ........................................................................................... 46

ANNEX 5. BORROWER COMMENTS ...................................................................................... 59

ANNEX 6. SUPPORTING DOCUMENTS .................................................................................. 69

ANNEX 7. TABLES AND FIGURES ........................................................................................... 70

The World Bank ACCESS WITH QUALITY TO HIGHER EDUCATION PROJECT-ACCES II -SOP PHASE 2 (P145782)

Page 1 of 73

DATA SHEET

BASIC INFORMATION

Product Information

Project ID Project Name

P145782 ACCESS WITH QUALITY TO HIGHER EDUCATION PROJECT-

ACCES II -SOP PHASE 2

Country Financing Instrument

Colombia Investment Project Financing

Original EA Category Revised EA Category

Not Required (C) Not Required (C)

Organizations

Borrower Implementing Agency

ICETEX with the guarantee of the Republic of

Colombia

Instituto Colombiano de Crédito Educativo y Estudios

Técnicos en El Exterior (ICETEX)

Project Development Objective (PDO) Original PDO

The objective of the Project is to increase student enrollment, graduation and equity in higher education, by: (a) increasing the number of ACCES Student Loans and the ACCES Program’s focus on students from disadvantaged socioeconomic backgrounds and in quality higher education institutions and programs; and (b) enhancing ICETEX’s institutional capacity.


Page 2 of 73

FINANCING

Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$)

World Bank Financing IBRD-83540

200,000,000 200,000,000 200,000,000

Total 200,000,000 200,000,000 200,000,000

Non-World Bank Financing 0 0 0

Borrower/Recipient 236,000,000 0 0

Total 236,000,000 0 0

Total Project Cost 436,000,000 200,000,000 200,000,000

KEY DATES

Approval Effectiveness MTR Review Original Closing Actual Closing

01-Apr-2014 28-Jan-2015 30-Jun-2019 15-Mar-2019

RESTRUCTURING AND/OR ADDITIONAL FINANCING

Date(s) Amount Disbursed (US$M) Key Revisions

13-Mar-2019 200.00 Change in Loan Closing Date(s)

KEY RATINGS

Outcome Bank Performance M&E Quality

Satisfactory Satisfactory Substantial

RATINGS OF PROJECT PERFORMANCE IN ISRs

No. Date ISR Archived DO Rating IP Rating Actual

Disbursements (US$M)

01 12-Jul-2014 Satisfactory Satisfactory 0

02 29-Jan-2015 Satisfactory Satisfactory 0

03 01-Nov-2015 Satisfactory Satisfactory 59.54


Page 3 of 73

04 09-Jun-2016 Satisfactory Satisfactory 75.30

05 20-Dec-2016 Satisfactory Satisfactory 120.90

06 29-Jun-2017 Satisfactory Satisfactory 200.00

07 09-Jan-2018 Moderately Satisfactory Moderately Satisfactory 200.00

08 16-Sep-2018 Moderately Satisfactory Moderately Satisfactory 200.00

SECTORS AND THEMES

Sectors

Major Sector/Sector (%)

Education 100

Tertiary Education 100

Themes

Major Theme/ Theme (Level 2)/ Theme (Level 3) (%) Human Development and Gender 0

Education 92

Access to Education 23

Science and Technology 23

Teachers 23

Standards, Curriculum and Textbooks 23

Labor Market Policy and Programs 10

Labor Market Institutions 5

Active Labor Market Programs 5

ADM STAFF

Role At Approval At ICR

Regional Vice President: Hasan A. Tuluy Axel van Trotsenburg

Country Director: Gloria M. Grandolini Ulrich Zachau

Director: Luis Benveniste


Page 4 of 73

Practice Manager: Reema Nayar Emanuela Di Gropello

Task Team Leader(s): Janet K. Entwistle Pedro Cerdan-Infantes

ICR Contributing Author: Uriel Kejsefman


Page 5 of 73

I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES

A. CONTEXT AT APPRAISAL

Context 1. Country context. At appraisal, Colombia was a top performer in Latin America in economic growth and poverty reduction. A sound macroeconomic context, a rich natural endowment, and favorable external conditions had driven “a decade of impressive successes,” including high income growth and poverty reduction (World Bank 2015). During the 2000s, the Colombian economy experienced an average 4.4 percent gross domestic product (GDP) growth and 4.8 percent thereafter (2010–2014). Extreme poverty fell from 17.7 percent in 2002 to 8.1 percent in 2014, while total poverty (including moderate poverty) fell from 49.7 percent in 2002 to 28.5 percent in 2014, in the process, lifting 6.7 million people out of poverty. 2. Despite these achievements, inequality remained high. The Gini index for 2012 was 0.54, the country’s lowest level in a decade and yet one of the highest in the region. The richest 10 percent of the population possessed 42 percent of the nation’s income, while the bottom-40 percent (B-40) possessed only 10 percent. Inequality was especially noticeable among ethnic groups and between regions. While the Bogota region was responsible for 73 percent of output, remote areas of the country (zonas apartadas) had income levels that were a small fraction of urban centers. In response, the Government of Colombia (GoC) launched a National Development Plan (Plan Nacional de Desarrollo, PND) for the 2010–2014 period titled “Prosperity for All.” 3. Sector context. At appraisal, the key challenges for the education sectors were identified as “expanding enrollment and graduation rates for the overall population and for students throughout [Colombia] and from disadvantaged socioeconomic backgrounds, as well as increasing the number of accredited institutions” (World Bank 2014). These were consistent with the chief planning document for the education sector in Colombia, the National Decennial Education Plan (Plan Nacional Decenal de Educación, PNDE) 2006–2016, which was organized around the axes of access, retention, and quality, with a transversal emphasis on equity. 4. During the previous years, Colombia had achieved remarkable progress in enrollment. Gross higher education enrollment had increased by 40 percent between 2006 and 2012, adding over half a million new students in the process. Most of this growth was due to the expansion of the National Learning Service (Servicio Nacional de Aprendizaje, SENA), which for the most part offered short-term technical courses; Colombia remained below regional and Organisation for Economic Co-operation and Development (OECD) peers in higher education enrollment, especially for undergraduate university degrees. The PND had set a gross enrollment rate target of 50 percent for 2014. 5. Measures of internal efficiency, most noticeably graduation, had followed a similar expansion. The number of graduates had increased by more than 50 percent since 2006 (241,049 students graduated in 2012 compared to 158,561 in 2006), with graduation rates increasing from about 21 percent in 2006 to almost 34 percent in 2012. Once again, this growth had not pushed Colombia far enough toward its aspirational peers such as Chile (38 percent) or the OECD (38 percent).


Page 6 of 73

6. The final, and always critical challenge in Colombia, was equity. Colombia possessed a highly privatized higher education system which perpetuated and systematized some socioeconomic inequalities. Inequalities were especially noticeable at the regional level as the majority of quality institutions were in metropolitan areas. The growth in enrollment had allowed a new type of student to access higher education institutions (HEIs) en masse—students from lower socioeconomic strata, less developed regions, and vulnerable minorities. 7. At the time of appraisal, education in Colombia was a highly privatized system1 and public financing of higher education in Colombia had a large demand-side component, especially when compared to peers. The main instruments were student aid and loan programs financed through the Colombian Institute for Educational Credit and Technical Studies Abroad (Instituto Colombiano de Crédito Educativo y Estudios Técnicos en el Exterior, ICETEX). ICETEX has been providing scholarships and student loans since 1950 and is the only provider of long-term student loans (especially for vulnerable populations). ICETEX has been growing in importance alongside the boom of higher education in Colombia. In 2013, almost 10 percent of higher education students received financing from ICETEX. 8. Since 2002, and with continuous support from the World Bank, ICETEX has been providing long-term student loans based on shifting priorities that are reflected in its allocation formula,2 which prioritize student merit (measured through standardized examinations), socioeconomic status, region of origin, quality of the institution or program (primarily measured through accreditation status) and expected repayment ability. In particular, the World Bank Access with Quality to Higher Education (Acceso con Calidad a la Educación Superior, ACCES) Project was fundamental in this growth in size and importance of ICETEX. The ACCES Program allowed for a rapid expansion with a simultaneous focus on equity—between 2004 and 2012, the annual number of student loans through ACCES doubled, and by 2012, 82 percent of student loans went to students from the lowest socioeconomic strata (compared to 30 percent in 2002). 9. Recognizing the long-term impact of the ICETEX agenda on outcomes such as salaries, employment, and the economy, the World Bank has been supporting the Colombian higher education agenda through ICETEX since 2002, through the Improving Access to Higher Education Loan (P074138, Loan 7155-CO), and in 2008, through the Second Student Loan Support Project Adaptable Program Loan (APL) Phase I (P105164, Loan 7515-CO). This project, Access with Quality to Higher Education Project – SOP Phase 2 (ACCES II), was prepared during 2013 and 2014, became effective on January 2015, and was launched on May 2015. It built on the World Bank’s longstanding engagement with the GoC and ICETEX, was guided by the priorities outlined in Colombia’s PNDE and PND and was closely aligned with the World Bank’s Country Partnership Strategy (CPS) for FY2012–FY2016 (Report No. 60620-CO) discussed by the Board on June 12, 2011, specifically in its call to expand opportunities for social prosperity and inclusive growth with enhanced productivity (see section on Relevance).

1 In 2013, 72 percent of all HEIs were private. 2 During project preparation, it was discussed that the loan allocation formula would be modified during the second semester of 2014, namely after project approval by the Board, with the dual objective of improving the ICETEX loan portfolio through a greater emphasis on lowering students’ default risk and on targeting vulnerable high-achievement students.


Page 7 of 73

Theory of Change (Results Chain)

Activities/Outputs Intermediate Outcomes PDOs/Outcomes

Expand tertiary education financing for vulnerable students

• Bank resources

• Student resources (partial co-financing and loan repayment)

• Government resources (subsidies)

• Partnerships (Alianzas, Administrative Funds, SER Scholarships)

• New sources (cuentas abandonadas, ICL, ahorro programado)

Increase the number of

ACCES Student Loans

Increase student

enrollment

Decrease students’ financial burden and barriers to entry

• Maintenance subsidies from the GoC

• Lower the required payments during period of studies and

increase loan maturity (ACCES, Tu Eliges)

• Eliminate the need for loan co-signer (Fondo de Garantías)

Lower the cost of

attending and

remaining in higher

education

Increase graduation

Prioritize accredited HEIs and programs and incentivize

accreditation

• Reform of loan allocation formula to favor enrollment in

accredited HEIs

Increase enrollment in

accredited HEIs and

programs

Prioritize high-achievement students

• Reform of loan allocation formula to favor high-achievement

students and those more likely to repay

Increase the proportion

of high-ability students

Decrease regional inequality in access to higher education

• Reforms to loan allocation formula to prioritize students from

Zonas Apartadas

• Dissemination in Zonas Apartadas (oficinas móviles)

Increase the number of

students from zonas

apartadas

Increase equity

Decrease socioeconomic inequality in access to higher education

• Increasing the ACCES Program’s focus on students from

disadvantaged socioeconomic backgrounds

• Better targeting of the poorest (SISBEN III)

Increase in the number

of students from

socioeconomically

vulnerable households

Decrease ethnic inequality in access to higher education

• Expansion of Administrative Funds

• Indigenous People’s Plan (IPP)

Increase in the number

of students from

vulnerable minorities

(Afro, Roma, Victims)

Improve ICETEX’s financial sustainability

• ICETEX patrimony expansion (financial innovation, better cost

recovery, lower administrative costs)

• High credit rating

• Repayment strategies (pre-judiciary management)

• Modernization of technology and collection software

• Diversification of funds

• Impact evaluations

Increase and improve

the management of

ICETEX’s financial

portfolio

Enhance ICETEX’s

Institutional

Strengthening

Note: ICL = Income-contingent loan; IPP = Indigenous Peoples Plan; SISBEN = Identification System for Social


Page 8 of 73

Program Beneficiaries (Sistema de Identificación de Beneficiarios de Programas Sociales).

10. The Project’s Theory of Change shows the links between project activities and key development outcomes of increasing student enrollment, graduation, and equity in higher education.

11. This Theory of Change belongs in part to a larger development strategy framed by a sequence of projects. It builds upon the Theory of Change of Phase I of the APL and through it, on the accumulated progress and lessons of a 17-year partnership between the World Bank and the GoC. This includes continuing existing goals with greater ambition (for example, Phase I raised enrollment numbers from 1.2 million in 2006 to 1.8 million in 2012, while ACCES raised these values to 2.32 million in 2018) and introducing changes that carry on to future projects (such as the institutional consolidation).

12. The institutional strengthening of ICETEX, a major component of the project, is an intermediate outcome to support other key development outcomes. A more transparent, efficient, and sustainable ICETEX can both leverage more independent resources from private markets and better use those available to maximize the development impact of the project. Institutional strength is ultimately the foundation upon which other development outcomes are achieved.

13. In the long term, the project contributes to the economic and social development of the country and the individual wellbeing of beneficiaries. Individuals who complete higher education receive, on average, significantly better salaries, more employment opportunities, and more employment choice. Together, these improvements sustain and expand growth.

Project Development Objectives (PDOs)

14. The objective of the project is to increase student enrollment, graduation, and equity in higher education, by (a) increasing the number of ACCES Student Loans and the ACCES Program’s focus on students from disadvantaged socioeconomic backgrounds and in quality HEIs and programs; and (b) enhancing ICETEX’s institutional capacity. The program objective of the Series of Projects (SOP) is the same as the PDO of this project.

Key Expected Outcomes and Outcome Indicators

15. The PDO was to be measured through the following four Project Development Objective Indicators (PDOIs):

(a) Enrollment in higher education (b) Graduation in higher education (c) Percentage of new students enrolled in higher education from families earning less than 2

minimum salaries (d) Percentage of graduates from families earning less than 2 minimum salaries

Components 16. The project financed the following two components:


Page 9 of 73

Component 1: Student Loans (Appraisal and Actual: total: US$424 million; World Bank: US$200 million; ICETEX: US$135 million; Government: US$89 million) 17. Component 1 sought to increase student enrollment and graduation and improve equity in higher education by increasing the number of new and renewed ACCES student loans provided by ICETEX, and by increasingly focusing them on students from socioeconomically disadvantaged backgrounds and quality HEls and programs. The focus on quality programs with higher returns in the labor market and lower default, would simultaneously improve the probability of loan repayment and strengthen ICETEX's financial soundness. 18. Component 1 funded ACCES student loans to eligible students to finance their tuition and/or maintenance (sostenimiento) in eligible HEls to pursue undergraduate degrees through both new loans and loan renewals for continuing students. Component 2: Institutional Strengthening (Appraisal and Actual: total: US$12 million; World Bank: US$0 million; ICETEX: US$12 million) 19. Component 2 sought to improve ICETEX’s management practices, help it diversify funds, and conduct an impact evaluation to support the project’s key objectives.

(a) Strengthening ICETEX's loan administration and portfolio management. This subcomponent sought to support the development of differentiated collection strategies based on portfolio risks, and the upgrading of information systems.

(b) Strengthening ICETEX's capacity to manage its ACCES Program and the Sustainability Fund. This subcomponent sought to support the use of scorecards and the carrying out of HEI contracts’ implementation.

(c) Provision of support for ICETEX's diversification of financial resources. This subcomponent sought to support the studies, including on alternatives to expand its funding sources and to develop an endowment fund, and on partnerships with the private sector and fundraising activities.

(d) Enhancing ICETEX’s knowledge base. This subcomponent sought to support the understanding and decision making of ICETEX by helping conduct impact evaluations of its student loans and tracer studies to track employment outcomes of ICETEX’s beneficiaries who have completed their studies and supporting ICETEX's overall project coordination, implementation, and monitoring (including audits).

B. SIGNIFICANT CHANGES DURING IMPLEMENTATION

Revised PDOs and Outcome Targets 20. The project PDOs or outcome targets were not revised during the life of the project. Revised PDO Indicators 21. The project PDOIs were not revised during the life of the project. Revised Components 22. The project components were not revised during the life of the project.


Page 10 of 73

Other Changes 23. The closing date was advanced by 3.5 months, from the original closing date of June 30, 2019 to the revised closing date of March 15, 2019. Rationale for Changes and Their Implication on the Original Theory of Change 24. The closing date was advanced to reflect the faster disbursement and implementation of the project. A number of developments, including a structural reform of the loan allocation system (the ‘Tu Eliges’ program, implemented in 2015), caused a growth in the average cost of each student loan. Tu Eliges drastically diminished the average student contribution during the period of studies, which was covered by ICETEX and thus World Bank resources (see ‘Key Factors for Implementation’ for a more comprehensive explanation). The proportion of financing covered by student resources was reduced dramatically and substituted by World Bank funds, which were therefore disbursed much faster. This change had no implication on the original Theory of Change; it only affected the speed of implementation. 25. As a result of this accelerated disbursement, the project’s last disbursement took place in January 2017, two and a half years ahead of closing. The last major activity of the project (the completion and delivery of the project’s Impact Evaluation) took place in May 2018, more than a year ahead of closing. Moreover, a subsequent operation of US$160 million (Access and Quality in Higher Education Project – PACES, P160446) was approved by the Board in January 2017; and an additional financing of another US$160 million for this operation became effective on January 2019. This new operation, also in partnership with the GoC and ICETEX for the higher education sector, guarantees that key outcomes and priorities of this project will continue to be monitored far beyond the original closing date (PACES’ closing date is December 1, 2023). Overall, with this project’s funding and activities fully completed and another operation taking the lead of supervision, the closing date was anticipated from its original June 30, 2019 to the revised March 15, 2019. 26. All PDO and PDOI indicators’ data is provided by the MEN but data for 2018 PDO indicators is not available yet. The project’s PDOI for enrollment and graduation are especially sensitive to time, due to the assumed natural growth rate of these that were included in the targets when they were set. Therefore, (a) for outputs and indicators under the direct scope and attributions of the project and its implementer (for example, carrying out an impact evaluation), the project is understood to finance an aggregate whole, which if disbursed early should still be met (in this case, the 2018–2019 end target); (b) for outcomes and higher-level results, especially national-level indicators for which the project and ICETEX contribute but do not determine, the project is expected to meet the yearly targets, in this case, 2018 targets or ‘Year 5’ in the Results Framework. II. OUTCOME

A. RELEVANCE OF PDOs

Assessment of Relevance of PDOs and Rating

27. Relevance of the project’s objectives is High. The project’s objectives were relevant to the CPS for FY2012–FY2016 during preparation (Report No. 60620-CO) and continued to be aligned throughout implementation. Tertiary education was a central component of the country strategy’s call to “expand opportunities for social prosperity” (strategic theme 1) through improved opportunities in education, and


Page 11 of 73

in its call to “promote inclusive growth with enhanced productivity” (strategic theme 3) through improved productivity and innovation (World Bank 2011). The CPS Progress Report for FY2012–FY2016 (Report No. 83966-CO) determined that the CPS continued to be relevant, adequately addressed the country’s development challenges, and addressed the priorities from the GoC, and further strengthened the alignment with the project by revising a CPS outcome indicator to match the project’s PDOI (World Bank 2014). Moreover, the project objectives continue to support the new Country Partnership Framework for the Republic of Colombia for FY2016–FY2021 (Report No. 101552-CO), discussed by the Board of Executive Directors on February 23, 2016, specifically through its second pillar of “enhancing social inclusion and mobility through improved service delivery” under its fourth objective of “improved access to and quality of education” (World Bank 2016). Finally, the project’s role in improving the salaries and the quality of jobs that beneficiaries can access with the long-term effect on sustainable economic growth and inclusion directly support the chief goals of the country’s 2015 Systematic Country Diagnostic (World Bank 2015). 28. During preparation, the project was also fully aligned with national development strategies. The ‘Prosperity for All’ 2010–2014 PND aligned with the project through its two strategic goals of fostering sustainable and competitive growth and creating equal opportunities for social prosperity. The PND noted the crucial importance of higher education to, among others, foster innovation and entrepreneurship, align competencies and jobs, and promote equity (Departamento Nacional de Planeacion 2010). The project became even more relevant for the ‘All for a New Country’ 2014–2018 PND, whose three key pillars (education, equity, and peace) are all being supported directly or indirectly by the project. Most relevant is the project’s impact on the education pillar through the strategic goal of making Colombia the most educated country in Latin America by 2025 (‘Colombia, la mas educada’). This goal is anchored in many of the project themes, such as closing access and quality gaps among socioeconomic and ethnic groups and region (objective a) or developing a tertiary education system with greater access, quality, and pertinence (objective d). This latter objective specifically notes the “expansion of student financing in a context of high quality and equity” (even naming ICETEX) as the key mechanisms for its achievement (Departamento Nacional de Planeacion 2014). It is thus clear that the project has supported the medium goals of the GoC at preparation and has been central to the Government’s own objectives and strategies in later documents. The signing of the peace agreement with the Revolutionary Armed Forces of Colombia (known as ‘FARCs’) in 2016 also added to the relevance of the project’s targeted support to the victims of the armed conflict. 29. Finally, relevance is further demonstrated by the consolidation of previous projects’ lessons and the rapid consolidation of the World Bank’s portfolio of support to education in Colombia. Two new operations with the GoC in education, the PACES Project (P160446) of 2017 and its additional financing (P166177) effective since 2019, have added US$320 million to “improve the quality of tertiary education in participating institutions and to increase the enrollment of students from disadvantaged socioeconomic backgrounds in quality programs” (World Bank 2017). In close alignment with the objectives of the ACCES Project, these new operations have continued the partnership with ICETEX and the focus on financing the demand with a focus on the most vulnerable and on high-quality institutions and programs. B. ACHIEVEMENT OF PDOs (EFFICACY) Assessment of Achievement of Each Objective/Outcome 30. To assess the project’s efficacy, this section is organized by the three key project objectives (enrollment, graduation, and equity). There are also two intermediate objectives (increasing the number


Page 12 of 73

of ACCES loans and enhancing ICETEX’s institutional capacity). Progress in increasing access loans is assessed within the efficacy analysis of the three key project objectives. The success in strengthening ICETEX is evaluated under the Institutional Strengthening section. This latter intermediate objective is especially central to the project’s Theory of Change (as illustrated by it having a specific component and the majority of PDOIs) but was not included directly as a key objective because the World Bank did not directly fund these activities. (a) Increase student enrollment 31. The project’s achievement in relation to increasing student enrollment is Substantial. Enrollment, as measured at the national-level by PDOI1, grew continuously during the life of the project and kept pace with medium-term goals of the GoC (as articulated in the PND). The number of students enrolled in tertiary education undergraduate university degree programs and professional technical and technological degrees (PDOI1) went from a baseline of 1.84 million students in 2012 to 2.24 million students in 2017 (latest available data), close to fully meeting the 2017 target (2.28 million), and substantially meeting the 2018 target (2.32 million). This represents a net gain of 400,000 students (or 22 percent) for the tertiary education system. Moreover, given the natural growth in national-level enrollment, the 2018 targets are likely to be met by the end of the corresponding year. During the life of the project, the gross enrollment rate increased from 42.4 percent in 2012 to 52.8 percent in 2017. 32. In addition, the project’s impact evaluations substantiate the direct contributions of ICETEX and the ACCES Program on increased student enrollment. Melguizo, Sanchez, and Velasco (2016) evaluate the impact of the ACCES Program through a regression discontinuity design and find that the ACCES Program improved by a remarkable 20 percentage points the probability that a beneficiary student would enroll in college. These results show the exceptional success of the ACCES Program’s design, especially when compared to more modest effect found in the literature on the impact of financial aid on enrollment (Fack and Grenet 2015), even in Latin America (Rau, Rojas and Urzua 2014). 33. Nonetheless, the overall number of ACCES loans increased less than the target of Intermediate Results Indicator 1 (IRI 1). During project preparation, ICETEX was allocating an average of 40,000 new ACCES loans yearly, which were expected to grow to 44,000 by 2018. Instead, the number of new ACCES loans decreased, reaching 20,266 in 2018. The number of loan renewals increased from 158,000 in 2012 to 186,072 in 2018, a net gain of more than 28,000 renewals. Overall, in 2018 ICETEX allocated 206,338 loans of the 262,000 that were expected or 79 percent of what the project anticipated for 2018. This represents an overall growth of more than 8,000 loans (a 4.2 percent increase from the baseline). 34. The main reason for a lower overall allocation of ACCES student loans is the aforementioned increase in the average costs of loans (Figure 1).3 The project was prepared assuming a semester cost-per loan of COP 2.5 million (US$862), while the actual cost in 2018 had reached COP 5.99 million (US$2,065), more than double the expected amount. This increase in the cost per student had three main sources: (a) the increase in the proportion of loans going to high quality institutions and programs recognized by the Government through the Accreditation, which are on average significantly more expensive, (b) the substantial growth in the proportion of the degree cost serviced by ICETEX with Tu Eliges, and (c) the growth in tuition costs across Colombia. The first supports quality while the second improves equity, both essential components of the project and of the type of enrollment that the project seeks to foster (one

3 Figures are found in Annex 7.


Page 13 of 73

leading to shared prosperity and economic growth). The focus on increased participation of loans to finance high quality education started with the inclusion of a Law in 2014-2018 PND that restricted ICETEX from lending to students in nonaccredited HEIs or programs. ICETEX immediately started increasing the proportion of loans going to accredited institutions and, by January 2018, only students attending accredited HEIs or programs were granted ICETEX’ loans. This change of focus, while costed the number of new loans and renewals to perform below end-of-project targets (46% and 85%, respectively); progressively increased results for high quality education and by 2018, 83% of ACCES loans were given to students attending accredited institution or programs, far above the final target of 40%. However, equity concerns emerged related to the lack of accredited institutions in remote areas and on July 2018, the Government abolished the restriction. 35. Overall, the ACCES Program managed to positively contribute to increasing enrollment as the national-level enrollment targets and the impact evaluation corroborate, while contributing to its development objective of increasing enrollment with quality and equity. Despite a rise in costs, ICETEX managed to deliver more quality loans than the project funds would allow and is considered to have substantially achieved its enrollment objective. (b) Increase student graduation 36. The project’s achievement in relation to increasing student graduation is Substantial. Graduation numbers, as measured at the national level by PDOI2, have grown during the life of the project from 241,000 in 2012 to 365,000 2017, an expansion of 124,000 graduates or 50 percent over the baseline, close to the 2017 target of 375,000. Further substantiating success toward this objective, the average graduation rate improved from 4.7 percent in 2001–2010 to 6.7 percent during the life of the project (2012–2017). The dropout rates continued to fall during this period (Figure 2). Most encouragingly, not only did dropouts decrease and graduation improve but the relevance of education increased—the rate of labor insertion in the formal sector has grown from 74.3 percent in 2012 to 76.5 percent in 2016. 37. Furthermore, the impact evaluation showed that “ACCES decreased dropout rates by about 7 percent and increased academic performance for the students at the margin” (Melguizo, Sanchez and Velasco 2016). Decreasing dropouts is a key factor in increasing graduation, thus serving as supporting evidence of the program’s positive impact on graduation. Similarly, student performance is a good predictor of dropout, and ACCES increased the percentage of college subjects passed or approved by 3–4 percent for students at the margin with respect to comparable students without ACCES loans (73 percent versus 76 percent). 38. The other key intervention of the project toward graduation, namely subsistence subsidies (which along with interest rate subsidies lower the cost of remaining in higher education) were highly effective. Lozano-Rojas (2018) showed that maintenance subsidies lower the probability of student dropout by 23.5 percentage points (from 39 percent to 16.5 percent for those students who get it) and increase the probability of graduating in six years by 12.8 percentage points (from 21.3 percent to 34.1 percent). In the first year alone, the subsistence subsidies reduce the probability of dropout by 15 percentage points. 39. Considering that at the national level, (a) the number of graduates has increased by 124,000 (only slightly below the 2017 target of 134,000), (b) labor market insertion of those graduates improved, (c) the dropout rate continued to fall and, (d) ACCES was shown to reduce the dropout rate by 7 percent and improve student academic performance, while maintenance subsidies increased graduation by 12.8


Page 14 of 73

percentage points, this objective is rated Substantial, as other than a slight underperformance in the number of graduates, it fully achieved its intended outcomes. (c) Increase equity 40. The project’s achievement in relation to increasing equity is Substantial. National-level equity numbers have shown substantive improvements. In terms of enrollment, the proportion of new higher education students from families earning less than two minimum salaries (PDOI 3) has gone from a baseline of 58.10 percent in 2012 to 64.37 percent in 2017, surpassing the expected targets for 2017 (128 percent achievement) and even the final target (101 percent achievement). 41. Moreover, better indicators of socioeconomic vulnerability paint an even more impressive picture. When measuring students by the country’s most sophisticated multivariable poverty index, SISBEN the poorest have clearly made the largest gains (Figure 3). Enrollment of new students in higher education from students of SISBEN level 1 (the lowest level) have gone from 23.49 percent in 2012 to 31.47 percent in 2016, a remarkable 7.98 percentage points increase (34 percent improvement). When accounting for students of SISBEN of all levels, improvement is 7.71 percentage points (15 percent). Similarly, enrollment progress can be measured by quintiles, which most directly fits the World Bank’s focus on the B-40. During the life of the project, the B-40 expanded its participation in higher education significantly (Figure 4), going from being 20.8 percent of total enrollment in 2013 to 27.3 percent in 2017, an impressive 6.5 percentage points (or 31 percent) increase. The poorest quintile proportionally grew the most in those years (41 percent), while the B-second quintile had the most substantive absolute gains (3.5 percentage points), while all other quintiles diminished their relative predominance in higher education. Indeed, during the life of the project, the gap between the richest and the poorest quintiles in terms of access to higher education shrank by 5.1 percentage points or 29 percent. 42. The project’s impact evaluations further confirm the impact on equity of the project. Melguizo, Sanchez, and Velasco (2016) show that ACCES loans have a larger positive impact on enrollment and academic performance (measured by the proportion of low-income students at the margin passing more courses) for students from the lowest socioeconomic strata. Because maintenance subsidies are only assigned to the poorest, their impressive impact in reducing overall dropout (23 percentage points) and increasing graduation (12.8 percentage points) is overall a net positive effect on equity (Lozano-Rojas 2018). This is further confirmed by the proportion of new ACCES loans allocated to students who qualify for a maintenance subsidy (IRI 3) going from a baseline of 29 percent to 80 percent in 2018 (a 850 percent achievement rate), and by the proportion of ACCES loans allocated to students who qualify for a maintenance subsidy and score below the SISBEN cutoff point (IRI 4), which went from a baseline of 22 percent to 68.2 percent in 2018 (a 770 percent achievement rate). This positive results of equity on graduation seem confirmed by PDOI 4 (achievement rate of 96 percent).4 43. The project also tackled equity through its support for vulnerable minorities, namely indigenous people, Afro-Colombians, Roms, and victims of the armed conflict. Here, the project succeeded far above expectations. In 2018, 49.9 percent of all ACCES loans were allocated to these vulnerable minorities, up from 9 percent at the project preparation stage. This 40.9 percentage points increase (1,636 percent achievement rate from what was anticipated) is especially impressive not only for the degree of progress

4 Ministry of Education stopped measuring this indicator, so the latest available data is for the year 2016, which is less helpful in making conclusive statements about the project’s current success in this dimension.


Page 15 of 73

on a crucial equity dimension and World Bank priority, but for what it meant for ICETEX. In 2016, this proportion was at 8.2 percent, below the yearly target. The World Bank team stressed this underperformance and worked with ICETEX on a strategy to remedy it. Due to the recommendations of the IPP, an aggressive dissemination campaign and more administrative funds allocated for these population, an impressive reversal took place as the number of beneficiaries from vulnerable minorities went from 3,070 in 2016 to 10,111 in 2018, primarily driven by increases in the number of indigenous people. 44. Despite these strong results for the equity of Colombian education at the aggregate level, equity within regions was less successful. The project recognized the centrality of regional inequity in Colombia and sought to address it by making students from remote areas (zonas apartadas) a chief equity focus. In this regard, the project did not perform as expected, with a proportion of students from zonas apartadas in 2018 at a lower level than during appraisal. As outlined in the project’s Implementation Status and Results Reports, ICETEX demonstrated a serious and credible commitment to equity, conducting an aggressive remedial campaign, analogous to the strategy used for vulnerable minorities. In this case, however, the strategy was less successful due to challenges not anticipated to their full extent at appraisal (a) the smaller universe of candidates for higher education (remote areas have far lower secondary education graduation rates, so enrollment growth as a proportion of the total is harder), (b) the very small supply of HEIs in these areas, especially accredited, which limits students ability to feasibly attend an HEI eligible for ICETEX funding (Figure 5), and (c) worse academic achievement for these populations (as measured by SABER 11, the high school exit exam). With respect to the baseline, the quality priority and national preference for loans going to accredited institutions and programs was especially damaging to students from remote areas. It must be noted that the ICR mission visited remote areas and found other challenges. Lack of information or even misinformation about ICETEX loans, administrative issues with the processing of loans, and lower average salaries after graduation (which reduce the returns of studying and an ICETEX loan) have combined to affect the reputation of ICETEX in many of these regions. With few exceptions, however, students overwhelmingly credited ICETEX for their ability to study (often as the only available mechanism) and advised other students to take ICETEX credits. 45. Overall, the project shows success in improving equity in education, especially socioeconomic and for students from vulnerable minorities. Even in the underperformance on regional inequality, ICETEX demonstrated commitment in securing progress across all forms of equity. Thus, as the project provided a meaningful contribution to reducing inequality in education in Colombia, this objective is rated Substantial. Justification of Overall Efficacy Rating 46. The three key dimensions of the PDO—enrollment, graduation, and equity—have improved significantly through the ACCES Program as shown by the impact evaluations and outcome and output indicators. While there were minor shortcomings, ICETEX repeatedly demonstrated a serious commitment to achieving the aim of the PDO (providing quality education equitably to as many Colombians as possible), so the overall efficacy is rated Substantial.


Page 16 of 73

C. EFFICIENCY

Assessment of Efficiency and Rating

47. The project’s efficiency is assessed as Substantial. An economic analysis of the first and second components of the project shows positive results in both the cost-benefit analysis and the Internal Economic Rate of Return (IERR). The economic returns of the project for society and each key stakeholder are consistently and substantially positive. Specifically, the project yielded an overall IERR for society of 13 percent. For key stakeholders, the returns were (a) for students: 5.3 percent to 14.5 percent, depending on the type of degree and the number of subsidies received; and (b) for ICETEX: 7 percent to 14 percent, depending on the type of student. These results are robust under various sensitivity scenarios and are corroborated by the project’s impact evaluation, whose own economic analysis found an IERR for the ACCES Program of 22 percent. Details of the economic analysis are included in annex 4.

48. The project’s final social IERR of 13 percent is slightly lower than estimations at the appraisal stage (IERR of 15 percent) but is within the range of traditional findings in the literature (around 12 percent for Latin America and 11 percent to 21 percent for Colombia, as found by Psacharoupulos and Patrinos 2004). It does not appear to be a substantial change in the returns to higher education investments for the society, although some reallocation of resources has occurred among stakeholders. The IERR for students was lower than at the appraisal stage due to the increase in the cost of attendance and the inclusion of social security contributions and income tax withholdings as a transfer from the students to the GoC. The IERR for ICETEX, on the other hand, improved substantially due to changes in the way that interest rates subsidies are calculated and the decrease in dropout rates that render better degrees of collection and, thus, better rates of return for the institution.

49. Second, this project had an extremely efficient execution. The project had an on-time signing and effectiveness, required no restructuring and had an accelerated disbursement and implementation. In line with the project’s objectives, Tu Eliges, successfully replaced most student contributions with loans, specifically helping students from lower socioeconomic backgrounds face less liquidity barriers to studying, and in return accelerating the project’s disbursement. Due to its full disbursement two years and a half ahead of schedule and its final implementation activity (second impact evaluation) taking place more than a year before the original closing date, the project was closed 3.5 months ahead of schedule. This early closing reflects both the client’s and the World Bank’s preference as the project was understood to have substantially met its objectives, especially as the new PACES operation took over the World Bank’s support to the education sector.

50. Finally, ICETEX itself underwent a significant improvement to its institutional efficiency indicators, a product of institutional strengthening efforts from Component 2. These included a reduction in the nonperforming loan ratio from 13.19 percent in 2014 to 8.17 percent in 2017 (Figure 6), and the development of valuable new financial products to consolidate the portfolio and sustainability of the institution.5 Overall, the high returns of the project, its accelerated implementation and its improvements to the implementer’s institutional efficiency suggest a Substantial achievement for the project’s efficiency.

5 See “financial innovations” under Institutional Strengthening.


Page 17 of 73

D. JUSTIFICATION OF OVERALL OUTCOME RATING 51. The relevance of the project’s objectives is High, efficacy is Substantial, and efficiency is Substantial. The overall outcome rating is therefore Satisfactory. E. OTHER OUTCOMES AND IMPACTS Gender 52. The project did not specifically seek to address the gender gaps in the education sector. Because females have a higher likelihood of graduating secondary education, they also have a higher probability of participating in higher education (52 percent during the life of the project) and of gaining an ACCES loan (54 percent at appraisal). The project tracked the percentage of female beneficiaries, starting at 57 percent and seeking to achieve a target of 54 percent, thus aiming to decrease the gender gap in favor of males. This bridging did not take place, and the gap increased slightly, with women representing 57.8 percent of beneficiaries in 2018. Part of the reason for the continuous increase in the gap between secondary education and higher education is that women have much lower dropout rates in tertiary education (46 percent compared to 54 percent for males). In this regard, the project’s focus on supporting academic progression and graduation may have helped males proportionally more, although concrete data is not available. Institutional Strengthening 53. Component 2 of the project was designed to support the institutional strengthening of ICETEX through several key activities to improve ICETEX’s long-term financial sustainability and independence. These activities ranged in objectives, from enhancing the institution’s knowledge base to modernizing its information technologies, or from improving its financial management (FM) to supporting the processing and administration of loans:

• Financial sustainability. During the life of the project, ICETEX improved a number of key indicators including (a) increasing its assets under management—both its overall assets and its cash flow increased (Figure 7), its external resources (IRI 14: surpassed), and its administrative funds, which target vulnerable minorities (IRI 15: surpassed); (b) keeping its administrative costs low (IRI 8: achieved); and (c) maintaining a good credit rating (Fitch: AAA rating maintained). An Economic and Financial Intermediary Assessment conducted in the preparation of PACES also confirmed the significant improvement in ICETEX’s financial situation during the past years, which confirmed the suitability of the institution to receive further funding.

• Financial innovation. ICETEX has developed new financial products that allowed it to maximize its available funds for students. These included (a) Cuentas abandonadas (see the following subsection on Mobilizing Private Sector Financing); (b) ICL: ICETEX led the approval of the regulatory framework for Colombia’s first experiment with ICL, a financing mechanism that has been greatly successful in other countries (for example, England and Australia) in financing higher education with greater equity—the first pilot is expected in 2020; (c) Ahorro Programado: ICETEX has been advancing the regulatory framework for families to open tax-incentivized savings account for their children’s university studies; and (d) SER Scholarships (see the following subsection Mobilizing Private Sector Financing).

• Administration of loans. ICETEX revamped its billing and collection procedures which helped it improve its loan performance ratios. Impressively, the nonperforming loan ratio went from 16.1 percent in 2012 to 8.17 percent in 2017. Activities included (a) direct management of prejudiciary


Page 18 of 73

billing, which helped provide personalized solutions and greater amiability in supporting students in default; (b) development of an online tool to monitor payment and control of billing for students in default (www.acuerdos.net); and (c) “financial regularization” workshops during 2016, 2017 and 2018 provided face-to-face support and debt refinancing for students in default.

• Information systems upgrade. ICETEX launched a comprehensive information technology (IT) strategy to improve its information technologies (IRI 9: achieved), chief among them being the first stages of implementation of a core banking software.

• Corporate governance. ICETEX led a project with the International Finance Corporation (IFC) that produced a diagnosis and an action plan to improve ICETEX’s corporate governance. The analysis reviewed the legal status of the institution, the structure and functioning of its Board, its transparency and dissemination policies, its risk and control policies, and so on. The report’s recommendations are being used to design amendments to ICETEX’s regulatory framework.

• HEI support. Universities are ICETEX’s chief partner after its student borrowers, so ICETEX conducts regular university visits in which it sets up a temporary office that supports students in the understanding and management of their loan (IRI 11: achieved).

54. Knowledge base enhancement. Despite multiple World Bank operations with ICETEX, limited rigorous assessment of the ACCES Program existed. Thus, the project supported the development and implementation of two impact evaluations (IRI 16: achieved), which are extensively featured in this ICR due to their usefulness. Mobilizing Private Sector Financing 55. In an effort to diversify its sources of funding, expand its ability to lend to students, and increase its long-term sustainability, ICETEX made significant strides in mobilizing private sector financing to advance its objectives and those of the project. The main activities in this regard were the following:

• Cuentas Abandonadas. ICETEX successfully lobbied for the sanctioning of Law 1777 of 2016, which allowed ICETEX to invest resources from private banks considered “abandoned accounts,” namely private checking or savings accounts where no financial movements have been made for more than three years. ICETEX was then allowed to pool these private funds from Colombian banks and invest them through an independent fund under its administrations. The returns from these funds (roughly COP 900 billion) have been used to finance more student loans and quality improvements for HEIs.

• SER Scholarships. This fund, still in the legal approval process, will collect donations from the private sector toward fostering higher education access in Colombia. ICETEX and the National Education Ministry (Ministerio de Educación Nacional, MEN) are processing tax benefits and incentives for private entities to participate in the fund. It is expected that this line of scholarships will be launched during the second semester of 2018.

Poverty Reduction and Shared Prosperity 56. The project’s exclusive focus on the socioeconomically vulnerable helped reduce poverty and improve shared prosperity in Colombia. The project provided student loans to roughly 168,000 new students and helped renew loans for another 990,000. All these students come from the lowest strata of the Colombian populations or belong to constitutionally protected minorities. Like in most countries, unequal access to higher education is a handicap to social mobility and a chief reason for intergenerational poverty. This is especially true for a country of very large higher education returns, such as Colombia

http://www.acuerdos.net/


Page 19 of 73

(Ferreyra, et al. 2017). Melguizo, Sanchez, and Velasco (2016) showed the higher impact of ACCES loans for poorer students as this aid significantly raises their probability of enrolling in higher education, and therefore, receiving higher returns over their lifetimes. Focus groups interviewed during the ICR mission equally asserted that without ICETEX loans, they would have not had the means to enroll in college and pursue their careers in medicine, law, and the like. Finally, Afro-Colombians, indigenous people and people from ‘remote areas,’ who receive preferential treatment under the project, are often the most socioeconomically underprivileged within Colombia. III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME A. KEY FACTORS DURING PREPARATION 57. The project was designed to support the education agenda of the GoC, most clearly articulated at the time through the PNDE, which set ambitious national goals for enrollment growth. These ambitious goals were in part fueled by reasonable expectations at the time about future oil prices (one of the Colombia’s economy’s most crucial pillars and high at the time), which in these past years were frustrated by a drastic drop. Still, the design of the project, focused on demand-driven financing, is sound and builds on the lessons of previous operations and Colombia’s highly privatized system. B. KEY FACTORS DURING IMPLEMENTATION 58. Various factors affected the implementation of the project. Chief among them was the implementation of the Tu Eliges allocation model for student loans, which began with the second semester of 2015. Spearheaded by a new ICETEX president, this allocation model changed, with respect to traditional ACCES loans, the eligibility requirements, the repayment terms, and most crucially, the proportion of total costs financed by ICETEX. Seven credit lines were offered, distinguished by eligibility requirements and identified most notably by the proportion of tuition costs paid by the student: zero percent, 10 percent, 25 percent, 30 percent, 40 percent, 60 percent, and 100 percent. Thus, in a zero percent Tu Eliges loan, ICETEX would lend 100 percent of tuition costs to the student. The World Bank only co-financed the credit lines directed toward vulnerable students (zero percent, 10 percent, 25 percent). The ability to choose among these credit lines (and to obtain Government subsidies) depends mainly on achievement in standardized examinations. The zero percent credit line was the most notable innovation, as traditional ACCES loans only financed 50–75 percent of tuition costs; this new credit line allowed students with no additional resources to borrow the entire amount needed for studying. The second key innovation was the Fund of Guarantees (or Fondo de Garantías), which allowed approved students to borrow even in the absence of a loan co-signer (ACCES loans required it). Together, these innovations of Tu Eliges would allow the very poorest, who often lack any family saving or collateral, to also access higher education. The other most consequential aspect of Tu Eliges was a significant increase in the average loan disbursement. Given the option of minimizing parental contribution, most students took this option and zero percent was the most requested credit line; at a minimum, this would entail an increase of 25–50 percent in the cost of these loans for ICETEX. Adding the progressive preference for students going to accredited degrees and programs (which on average are more expensive than non-accredited programs), and the average inflation of university tuition above average inflation,6 meant that

6 A comparison of the growth of tuition versus overall inflation shows that the cost of education has been slightly above


Page 20 of 73

the cost per student grew enormously during the life of the project (Figure 1). This sharp increase in cost per student would constrain the breadth of the project, that is, reduce the number of loans that ICETEX could allocate with its resources. 59. The second key development was the Government’s purposeful push for quality in higher education through ICETEX. The GoC had included quality as a key axis for improving education in its 2010–2014 PND (Departamento Nacional de Planeacion 2010). However, in its 2014–2018 PND (Law 1753), the GoC went further and restricted ICETEX from lending to students in nonaccredited HEIs from 2018 onward. The goal of this restriction, given ICETEX’s key role as higher education’s main lender, was to incentivize students to choose accredited institutions and to incentivize nonaccredited institutions to seek accreditation before the 2018 deadline. ICETEX immediately started increasing the proportion of loans going to accredited institutions; indeed, the project’s quality indicator (IRI 6) was surpassed. However, accreditation is a lengthy process and little progress from the supply side was made to advance in accreditation. As a result, the universe of eligible universities for ICETEX credits diminished substantially. This change had especially devastating effects on ICETEX’s lending in remote areas, where the accredited offerings are especially scarce; students in remote areas found themselves unable to study within their own regions. Because of this consequential spillover effect, ICETEX received approval to eliminate the implementation of this requirement, so since August 2018, ICETEX is once again able to finance nonaccredited institutions. This crucial elimination suggested that some lagging indicators (zonas apartadas, new ICETEX loans allocated) would increase to levels more closely aligned with the expected targets; yet, results will not reflect in 2018 as the requirement was eliminated late during the universities’ schedules. 60. The final key development was the important change in the health of public finances during the life of the project. The economy of Colombia suffered a blow with the oil price shock of the second half of 2014, widening the current account deficit, which led to a depreciation of the currency and most crucially reducing Government revenue by 3 percent of GDP (World Bank 2016). This decrease in Government resources came accompanied by two important increases in Government spending (a) the peace process, whose post-conflict required substantial additional expenditures over the medium term, and (b) the Ser Pilo Paga program, which yearly funds about 40,000 top academic students from vulnerable backgrounds to access higher education programs with condonable credits (more similar to a scholarship than a loan). Together, these changes put pressure on the public resources from the GoC that were originally destined for ICETEX. IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME A. QUALITY OF MONITORING AND EVALUATION (M&E) M&E Design 61. The project’s Results Framework was clearly articulated and aligned with the theory of change and intended outcomes. The PDOIs were linked to the PDO, which provided clarity for monitoring and were strongly reinforced by the conception and inclusion of rigorous impact evaluations. The PDOIs also

inflation, contributing to the increase in the average disbursement amount. Still, the premium is not high enough so far as to raise flags on a cycle of public lending and tuition growth (Bennet hypothesis). Monitoring that this remains the case will be important as ICETEX’s and the Government’s funding for education continue to grow.


Page 21 of 73

tried to focus on the target population aligned with Government policy using socioeconomic strata, SISBEN, and minimum salaries, which are more precise and updated measurements of socioeconomic inequality in Colombia. On the other hand, the PDOIs focus on national-level enrollment and graduations figures was beyond the direct range of influence of ICETEX (as ICETEX covers a tenth of that population). Still, this shortcoming was mitigated by the inclusion of rigorous impact evaluations in the project and in the Results Framework. These impact evaluations further provided significant evidence of the project’s value-added and corroborated the results of the other indicators. IRIs were narrow and very closely aligned with the scope of ICETEX’s attributions and relied on ICETEX’s own data which was always readily and timely available. M&E Implementation 62. ICETEX proved great competency in regularly measuring and evaluating the progress of project activities during implementation. Implementation support missions reported timely data and MEN provided timely delivery of biannual reports. ICETEX supported the World Bank’s supervision missions, including visits to the field, that allowed the collection of additional valuable data. Regular surveys and open channels to express grievances have given necessary room for beneficiary feedback throughout the project’s implementation. M&E Utilization 63. The project did an effective job of utilizing the available data and knowledge produced. ICETEX showed true commitment to learning and improving from the lessons of this monitoring, as was demonstrated by the remedial activities to improve results of vulnerable minorities and students from remote areas, once the underwhelming results were received. What worked for vulnerable minorities and somewhat failed, for students from zonas apartadas does not change the fact that from focus groups across the country, ‘ICETEX mobile offices’, and targeted credit lines, ICETEX used the knowledge to create remedial actions. 64. Moreover, the implementation of the project’s impact evaluation was appropriate and was responsible for the quality product delivered. It also helped improve the evaluation capacity of ICETEX, which hired consultancies with IFC and other private consultancies (Genesis) during these past years to help them improve their operations and monitoring abilities. Justification of Overall Rating of Quality of M&E 65. The overall design and implementation of the project’s M&E system was suitable, the project’s impact evaluations were of substantive value, and the project implementers did proper collection and reporting of data and a meaningful use of the system to create tangible improvements to the project. Because the scope (not the concept) of PDOIs was the only minor shortcoming, the overall quality of M&E is rated Substantial. B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE

66. Social safeguards. The implementation of the Operational Policy (OP 4.10) of indigenous peoples for the ACCES Project is triggered as a mechanism of protection and safeguard to ICETEX beneficiaries, with a focus on inclusiveness across gender, culture, and on an intergenerational dimension.


Page 22 of 73

67. The supervision carried out for social safeguards aspects of the project show a progressive effort to improve user-assistance mechanisms, dissemination of information, and targeting of potential beneficiary populations, especially in remote areas of the country (Guajira, Amazonas, Arauca, Caquetá, Guainía, Guaviare, Chocó, Putumayo, San Andrés, Vichada, and Vaupés) where indicators remain below target. 68. The project supported improvements to enrollment and permanence strategies for vulnerable populations and pertinent evaluation instruments such as the development of Affirmative Action Plans with HEIs, focus groups with beneficiaries, and verification of the grievances and customer service system. 69. Environmental safeguards. The project is classified as category C. It does not include construction or rehabilitation of infrastructure; therefore, the OP/BP 4.01 on Environmental Assessment is not triggered. 70. Procurement. The project had no World Bank-financed procurement. The World Bank financing of Component 1 (Student Loans) was not subject to the World Bank’s procurement guidelines, and all contracts for procurement of goods and services were completely financed by ICETEX. No exceptional requests to the World Bank for procurement of goods and/or services financed with World Bank proceeds were submitted. 71. Financial management. The project had adequate FM arrangements, which provided reasonable assurance that loan proceeds were used for the intended purposes and FM requirements were complied with throughout the implementation period. ICETEX, as the implementing entity, was in charge of all FM aspects. Project accounting was integrated within ICETEX’s APOTEOSYS accounting system, where budgeting, accounting, and treasury modules are interfaced. Project records were updated, and accounts duly reconciled. Furthermore, there was an adequate control framework for the project, which included internal reviews performed by the Office of Internal Control; also, the entity was audited by the Office of the Comptroller General of Colombia. Interim Financial Reports, on a biannual basis, were reported to the World Bank in a satisfactory manner. The auditors issued unmodified opinions, and these were acceptable to the World Bank. FM performance was consistently Satisfactory and FM requirements were complied with throughout the implementation period. C. BANK PERFORMANCE

Quality at Entry 72. The World Bank assisted the GoC in the preparation of a relevant and implementable project, suited to the needs and priorities of the Colombian context, leveraging and deepening the partnership with ICETEX established since 2002 and designed with the vision that allowed a smooth implementation and an early completion. The project built on the successes and lessons from the Second Student Loan Support Project APL Phase I (such as transferring the responsibility for implementation from a Project Implementation Unit to ICETEX or maintaining the financial terms of the World Bank loan), while simultaneously pushing ICETEX to prioritize its institutional consolidation, funds diversification, and portfolio risk reduction. The World Bank also built a clear theory of change and adequate indicators to monitor it, including objectives clearly expressed upfront and a commitment to developing strong impact evaluations. The procurement and financial arrangements were clear, simple, and instrumental in allowing for the swift disbursement of the project. Finally, the World Bank did a correct assessment and


Page 23 of 73

mitigation of risks. While the accelerated disbursement of project funds suggest that these funds were insufficient in their conception, this miscalculation was mitigated by the strength of the partnership built with ICETEX, which allowed the new PACES operation to inherit and continue the project’s mission. Quality of Supervision 73. The project’s successful implementation (timely effectiveness, early disbursement, and implementation) was anchored in strong World Bank supervision. The World Bank conducted regular supervision missions and produced clear progress reports, valuable feedback to counterparts, and technical assistance to keep the project on track and under satisfactory progress. The continuous monitoring and support of the issue of vulnerable minorities and zonas apartadas illustrates the positive value added of World Bank supervision. 74. The project also experienced no major changes or implementation issues. Neither financial management nor procurement, nor safeguards experienced any hiccups, and the project needed no restructuring to accomplish a complete implementation. One World Bank team task leader transition and two leadership transitions at ICETEX did not disturb implementation. Moreover, during implementation, the partnership with ICETEX continued to flourish, as shown for example by the continuation through the PACES Project. Justification of Overall Rating of Bank Performance 75. Given the accelerated disbursement and implementation of the project, and only minor shortcomings in design, the overall World Bank performance is considered Satisfactory. D. RISK TO DEVELOPMENT OUTCOME

76. The development outcomes sought and achieved in the project are likely to continue. The partnership with ICETEX is secured by the PACES Project, whose approved additional financing guarantee that had a more ambitious set of objectives, adequate availability of resources (US$320 million), and World Bank technical assistance will continue to guide the World Bank’s engagement to the education sector in Colombia. 77. Still, the overall context presents a few challenges and unknowns. First, the GoC has changed both in its leadership and political tone and ICETEX itself has embarked in a process of institutional transformation which is expected to be in place on September 2019. This change opens the possibility that a change of priorities takes place with the new administration. Foreseeing this situation, the World Bank team in Colombia is working closely with ICETEX to provide technical assistance and accompaniment within the transformation process as a mitigation measure. Second, as one the project’s impact evaluations demonstrates, the loan subsidies of the GoC are essential to the viability and success of the project’s objectives (Lozano-Rojas 2018). The amount allocated to ICETEX student loans for subsidies has remained steady during the life of the project (financing roughly 20,000 students), but is renewed and renegotiated every year, and as such, is reliant on the health of public finances. The suspension or substantial reduction of these subsidies would jeopardize the accomplishment of project objectives. Third, if the mounting cost per student loan continues to rise, it would further endanger ICETEX’s lending capacity and overall enrollment.


Page 24 of 73

78. These risks, while real, are tempered by some attenuating circumstances. The World Bank’s engagement with ICETEX has received support across political parties and governments since 2002. Moreover, while ICETEX was a topic of debate during the election cycle, the current administration was among the most favorable to continue with current policies. As such, neither subsidies nor the project design are likely to change substantially. With regards to the cost per student, ICETEX recovered the ability to lend to nonaccredited institutions in July 2018, which should bring down the cost per loan. 79. Overall, given the continuation of World Bank engagement through PACES and the aforementioned attenuating circumstances, the risk that development outcomes achieved under this project might not be sustained, is considered Low. V. LESSONS AND RECOMMENDATIONS

80. Financing student loans for underserved populations is sound public policy. The project and its rigorous evaluations demonstrate empirically that well-designed, demand-driven financing mechanisms such as ACCES can lead to significant positive development outcomes in education enrollment, equity, graduation, and quality, providing privately life-changing and socially cost-effective returns.

81. ICETEX should critically review and drastically improve its operational processes, especially in the regions. Focus groups with students revealed meaningful shortcomings in the delivery of daily processes, especially outside the big metropolitan areas. For example, students sometimes receive their disbursements late, the paperwork required is unknown or cumbersome to process, tracking of debts and repayment options are confusing, and in-person representatives are insufficient in number and quality. ICETEX has still a long way to go in simplifying its bureaucratic requirements, enhancing the quality of its representatives, and ensuring that payments to institutions and students are delivered on time.

82. The very poorest and students from remote areas need better-targeted remedial activities. While ICETEX has meaningfully helped improve education opportunities for vulnerable students, the very poorest (strata 1) or those from remote areas have made relatively fewer gains. These students represent a particularly difficult institutional challenge even among the best achievers and will necessitate more robust and strategic interventions. Among them are remedial academic programs and coaching, incentives for students and institutions toward graduation, targeted financial products, and support to regional institutions for accreditation.

83. ICETEX should continue to revamp its external financial resources to become independent. After decades of support from the World Bank, ICETEX has become a more financially sustainable institution, with various new sources of funds, lower portfolio risk, and access to private markets. Consolidating this progress will be essential to achieve financial independence and maximize its development impact.

84. Enhancing the financial literacy of loan recipients is critical for institutional transparency and reputation. The majority of students receiving student loans in Colombia have inexcusably weak understanding of the financial contract (and lifelong consequences) they are taking upon themselves. ICETEX’s attempts to remedy this knowledge gap have so far been insufficient. Current technologies allow for far more robust interventions. ICETEX should at least make it compulsory for prospective loan recipients to attend rigorous live or virtual financial education courses, which include at a minimum, the


Page 25 of 73

basics of taking on a loan, interest rates, returns to education across programs, financial calculator of debt, and a knowledge assessment.

85. Addressing ICETEX’s reputational issues will be essential to safeguard its sustainability. ICETEX continues to have significant reputational issues, many inherited from times when student loans were far more expensive and harder to repay, but many from the problematic processing issues and poor experience of current students. Due to these issues, ICETEX is often portrayed negatively in the media, protests of alumni are recurrent, and prospective beneficiaries shun away from its products as a precaution. Many of the recent changes, such as Tu Eliges or ICL, are steps in the right direction. Still, more is needed. Lowering interest rates from better administrative and FM, and more realistic lending and repayment expectations are needed to decrease the number of unsatisfied beneficiaries.

86. Government subsidies and innovations (such as the lack of loan co-signer) make a defining difference for the poorest. The most vulnerable are highly sensitive to liquidity constraints to entry to higher education, such as tuition costs to be paid during the period of studies, subsistence costs, or the need for a co-signer as collateral. Innovations and subsidies that reduce these barriers to entry are effective policy tools to incentivize greater academic enrollment and progression and can be greatly cost-effective. Their continuation is essential for the sustainability of achieved progress.

87. Careful and gradual calibration of a country’s accreditation system when it is the chief policy arm for quality enhancement is critical. Quality is imperative for any successful higher education policy, but it is also expensive. In light of this, future projects should explicitly include quality as a key development objective and tailor targets with careful consideration of the costs involved. In Colombia, the 2015 push for accreditation, while well-intended, was too ambitious and swift for a process that is costly to change and thus slow to improve. A more gradual adjustment remains the more realistic approach. Also, as revealed by the ICR mission, HEIs in Colombia struggle with the financial costs of accreditation rather than the will to change; thus, financing accreditation-conducing investments may be a more effective strategy to improve quality than incentives.


Page 26 of 73

ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS

A. RESULTS INDICATORS A.1 PDO Indicators Objective/Outcome: Increase student enrollment in higher education

Indicator Name Unit of Measure Baseline Original Target Formally Revised

Target

Actual Achieved at Completion

Increase student enrollment in higher education- 1. Enrollment in higher education.

Number 1.84 2.32 2.32 2.24

31-Dec-2012 30-Jun-2019 30-Jun-2019 29-Dec-2017

Comments (achievements against targets):

Target nearly achieved (91%) for 2017 and substantially achieved (83%) for 2018. This indicator was slightly behind its yearly targets for a while but has managed to bridge the gap. As of 2017, the enrollment indicator was close to full achievement and suggests that the final enrollment target will be met in 2018 by no less than 91% and will most likely be at or above target. As with all PDO indicators, the data is provided by the National Education Ministry (Ministerio de Educación Nacional, MEN) and were not available for 2018 at the finalization of this document.

Objective/Outcome: Increase graduation in higher education


Page 27 of 73


Target


Increase graduation in higher education- 2. Graduation in higher education.

Number 241000.00 406000.00 406000.00 365000.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 29-Dec-2017


Target nearly achieved (93%) for 2017 and substantially achieved (75%) for 2018. This indicator shows substantial progress (124,000 additional graduates than in 2012), but is 10,000 graduates short of the target for 2017. This is consistent with previous progress (10,000 below target most years) and suggests a similar degree of achievement by Project end (93-95%).

Objective/Outcome: Increase equity in higher education


Target


Increase equity in higher education- 3. Percentage of new students enrolled in higher education from families earning less than 2 minimum salaries.

Percentage 58.10 64.30 64.30 64.37

31-Dec-2012 30-Jun-2019 30-Jun-2019 29-Dec-2017


Target surpassed (128%) for 2017 and fully achieved (101%) for 2018.


Page 28 of 73

These results include higher education students in the National Learning Service (Servicio Nacional de Aprendizaje, SENA), slightly above the Project's target for 2017 and suggests a similar degree of achievement by Project end (100%).


Target


Increase equity in higher education- 4. Percentage of graduates from families earning less than 2 minimum salaries.

Percentage 49.40 59.00 59.00 54.78

31-Dec-2012 30-Jun-2019 30-Jun-2019 30-Dec-2016


Target substantially achieved (96%) for 2016. 2016 Target was 55%; achievement against yearly target was 96%. Final Target was 59%; achievement against final target was 56%. As with all PDO indicators, this indicator was calculated by the MEN, but information for years 2017 and 2018 is not available yet. This indicator is uniquely produced by crossing data from the National Information System for Higher Education (Sistema Nacional de Información de la Educacion Superior, SNIES) and the System for the Prevention of Dropouts in Higher Education (Sistema para la Prevención de la Deserción de la Educacion Superior, SPADIES), which is undergoing a major redesign and data migration process and caused the MEN to stop measuring the indicator since 2016. As a result, measuring this indicator with the current data necessitates an adaptation so it becomes comparable to previous iterations and thus consistent with the baseline and Project preparation targets. Upon our request, the MEN has tried to reproduce the indicator but has not managed to do so successfully and claims the results are not comparable. The latest comparable information then is from 2016, when the indicator was met at 96%; this information however seems insufficient to infer conclusively about the degree of final progress.


Page 29 of 73

A.2 Intermediate Results Indicators

Component: Component 1: Student Loans


Target


1. Number of new students obtaining an ACCES student Loan.

Number 40000.00 44000.00 44000.00 20266.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target significantly not achieved for 2018. The latest value is significantly below the yearly target. The gap from the targets has been increasing in the past years. The chief reason for the gap is the exponential growth in the average cost of student loans, especially since the "Tu Eliges" program, which has meaningfully reduced the number of loans that ICETEX can allocate with the same funds.


Target


2. Number of ACCES student loan renewals.

Number 158000.00 218000.00 218000.00 186072.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target not achieved (47%) for 2018.


Page 30 of 73

The latest value is slightly below the yearly target (2018) and the previous year's target (2017). Despite meaningful progress (28,072 additional renewals from the baseline), this indicator is below target (31,928 renewals). The chief reason for the gap is the exponential growth in the average cost of student loans, especially since the "Tu Eliges" program, which has meaningfully reduced the number of loans that ICETEX can renew with the same funds.


Target


3. Percentage of new ACCES student loans given to students who qualify for a government subsistence grant.

Percentage 29.00 35.00 35.00 80.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (850%) for 2018. The number of loans given with government subsistence grants has increased, significantly increasing the proportion with these grants, despite the fact that the total number of loans has decreased. The combined effect has been a more than doubling in the proportion of loans with subsistence grants from the baseline, multiple times the expected progress.


Target


4. Percentage of new ACCES Percentage 22.00 28.00 28.00 68.20


Page 31 of 73

student loans given to students who qualify for a government subsistence grant and score below the SISBEN cutoff points.

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (770%) for 2018. The "Tu Eliges" program has meant even more emphasis on the most socioeconomically vulnerable, and as the total number of loans decreases, the proportion of the poorest receiving ICETEX loans has increased substantially, more than twice the baseline value.


Target


5. Percentage of new ACCES student loans given to indigenous peoples, afro-Colombians, Roma and victims of violence.

Percentage 9.00 11.50 11.50 49.90

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (1,636%) for 2018. ICETEX made a significant effort to support vulnerable populations, achieving a remarkable 5.5x increase from the baseline in the proportion of loans going to them. The most benefited minority has been indigenous peoples (46.7% of loans to vulnerable minorities), in a positive sign


Page 32 of 73

of adherence to the Safeguards Plan and Indigenous Peoples Plan. Most impressive is the fact that this indicator was below target in 2016, but an intensive dissemination and prioritization effort from ICETEX managed to drastically reverse the trend.


Target


6. Percentage of new ACCES student loans given to students enrolled in accredited higher education institutions or programs.

Percentage 28.00 40.00 40.00 83.30

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (461%) for 2018. The preference in loan allocation for quality institutions has been a growing priority for the Government of Colombia and ICETEX. Most of the progress achieved was due to changes in the allocation formula that encouraged participation in accredited institutions. From January to August 2018, ICETEX was required by law to only finance loans for students going to accredited institutions or programs, which dramatically increased this indicator's performance (close to 100%).


Target


7. Geographic distribution of ACCES loans to new students in zonas apartadas.

Percentage 4.00 6.50 6.50 2.90

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Page 33 of 73


Target not achieved. ICETEX has maintained a policy by which any qualifying student from 'zonas apartadas' who applies for any ICETEX loan is granted the loan. During the life of the Project, ICETEX also conducted numerous dissemination campaigns on its loan offerings, including a "mobile office" that traveled across the country. Despite this "100% coverage" of students from remote areas, targeted subsidies, and the dissemination campaign, the scarcity of higher education institutions (especially of accredited institutions) and lower rates of secondary education completion in remote areas has made reaching the proportion of these students remain as a challenge.

Component: Component 2: Institutional Strenghtening


Target


8. Subcomponent 2.1--Annual administrative cost of ICETEX.

Percentage 3.10 3.00 3.00 1.90

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target achieved. The indicator was considered achieved if equal to or below to the end target. ICETEX managed to decrease its administrative costs as expected, and is now below its baseline of 3.1%.

Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at


Page 34 of 73

Target Completion

9. Subcomponent 2.1--Information systems upgraded.

Text C&CTEX, APOTEOSYS and CRM in use and networked to personal computers.

Implementation of the IT strategy

Implementation of the IT strategy

Implementation of the IT Strategy Completed

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target achieved. ICETEX improved its technology systems through the implementation of the core banking system, which should allow for more agile internal operations, better information management, and more responsive service to students.


Target


10. Sucomponent 2.1--Percentage of nonperforming loans

Percentage 36.00 19.00 19.00 17.05

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (111%) for 2018. ICETEX achieved a substantial reduction in the proportion of non-performing loans (19 percentage points). This was achieved through a


Page 35 of 73

package of activities that included a strengthening of pre-judiciary debt collection strategies, improvements in payments software, debt refinancing workshops, and outsourcing of some collection procedures.


Target


11. Subcomponent 2.2--Number of HEI Support Offices monitored by ICETEX through on-site visits.

Number 90.00 90.00 90.00 178.00

30-Dec-2014 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target achieved. The indicator was considered achieved if equal to or above the end target. ICETEX achieved and maintained throughout the Project's life the expected number of HEI visits and almost double the number of them in 2018.


Target


12. Subcomponent 2.2--Percentage of HEIs that perform above the benchmark on at least 5 of the 6 indicators in the scorecards.

Percentage 54.00 64.00 64.00 45.30

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Page 36 of 73


Target not achieved. ICETEX provided three justifications for the limited progress: (i) a significant increase in the benchmark for one of the variables (percentage of collection) with regards to the baseline; (ii) the universe of HEI increased (from 172 to 211) with the entrance of many low-performing institutions that have driven the average down; and (iii) the growth in HEI with accreditation was significantly below what was expected.


Target


13. Subcomponent 2.2--Sustainability index - Student loan attrition (non-renewal) and associated default rate among ACCES recipients.

Percentage 6.60 6.00 6.00 3.30

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (550%) for 2018. Default rates and non-renewals have decreased significantly and more than expected, indicating a strong improvement in the progression of ACCES students. This is due to better student selection mechanisms (for example, likely-student-default in the allocation formula), better student collection strategies (as outlined in the non-performing loans indicator), and more reasonable financial terms under "Tu Eliges."

Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at Completion


Page 37 of 73

Target

14 a) Subcomponent 2.3--Total sum of resources available for student credit generated from external sources (Funds under Administration).

Number 84.00 176.00 176.00 4947.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed (5,286%) for 2018. The massive increase in funds under administration (growing more than six times the baseline) significantly expanded the resources ICETEX has at its disposal, and thus its ability to support students and its long-term sustainability, while also supporting the large improvements in some equity indicators (many of these funds have supported vulnerable minorities). As a result, ICETEX did not allocate funding for these loans in 2018 and all funding came from administration funds.


Target


14 b) Subcomponent 2.3 --Total sum of resources available for student credit generated from external sources (ICETEX budget).

Number 96.00 88.00 88.00 0.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018



Page 38 of 73

Target achieved for 2017 and not achieved for 2018. All funding for student loans came from administration Funds in 2018. However, the total amount of resources for student loans rose from US$180 billion (baseline) in 2012 to US$4,947 billion in 2018, significantly surpassing the target of US$264 billion for the end of the project. However, as this indicator aimed to progressively increase the amount of resources ICETEX devoted to the Project from its own budget, results are considered below the expected target.


Target


15 a) Subcomponent 2.3--Number of new student loans given to indigenous peoples through ICETEX’s Vice-presidency for Administered Funds (VFA).

Number 1700.00 1000.00 1000.00 3424.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed. The indicator was considered achieved if equal to or above the end target. The number of loans to indigenous peoples was higher than expected by almost double the baseline and surpassed the end target by 342%.


Target



Page 39 of 73

15 b) Subcomponent 2.3--Number of new student loans given to Afro-Colombians through ICETEX’s Vice-presidency for Administered Funds (VFA).

Number 2300.00 1500.00 1500.00 3749.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target surpassed. The indicator was considered achieved if equal to or above the end target. By 2018, the number of new student loans to afro-Colombians was above the target (249.9%) and considerably above the baseline by almost 1.500 new students.


Target


16. Subcomponent 2.4--Impact evaluations (IEs) of ICETEX student loan programs have been designed and completed.

Text 2009 IE published in 2010

and 2013 IE underway to be published in 2014

2nd IE completed 2nd IE completed 2nd IE completed

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


Target substantially achieved.


Page 40 of 73

The Project managed to complete two high-quality impact evaluations: (i) Melguizo et al. (2016) and (ii) Lozano-Rojas (2018). These impact evaluations were not only designed and implemented successfully, but their content was highly valuable to understand and assess the effects of the Project.

Unlinked Indicators


Target


Direct project beneficiaries Number 198000.00 262000.00 262000.00 283940.00

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018

Female beneficiaries Percentage 57.00 54.00 54.00 57.80

31-Dec-2012 30-Jun-2019 30-Jun-2019 31-Dec-2018


The total number of Project beneficiaries is above what was expected by roughly 22,000 persons.

B. KEY OUTPUTS BY COMPONENT

Objective 1: Increase student enrollment in higher education

Outcome Indicators 1. Enrollment in higher education


Page 41 of 73

Intermediate Results Indicators 1. Number of new students obtaining an ACCES student loan 2. Number of ACCES student loan renewals

Key Outputs by Component (linked to the achievement of the Objective/Outcome 1)

1. 400,000 new students in the tertiary education system 2. 992,000 renewals of student loans 3. 168,000 new ACCES loans 4. ACCES increases by 20 pp the probability of beneficiaries enrolling in college

Objective 2: Increase graduation in higher education

Outcome Indicators 1. Graduation in higher education

Intermediate Results Indicators

1. Percentage of graduates from families earning less than 2 minimum salaries 2. Percentage of new ACCES student loans given to students enrolled in accredited higher education institutions or program 3. Number of accredited institutions


1. 124,000 new graduates in higher education 2. Maintenance subsidies increased graduation by 12.8 pp 3. Maintenance subsidies lowered dropout rates by 23.5 pp 4. ACCES decreased dropout rates by 7% 5. 55 pp increase in the proportion of students in accredited programs 6. 2.2 pp increase in the rate of labor insertion 7. 3–4% increase in academic performance for students at the margin

Objective 3: Increase equity in higher education

Outcome Indicators 1. Percentage of new students enrolled in higher education from families earning less than 2 minimum salaries


Page 42 of 73

2. Percentage of graduates from families earning less than 2 minimum salaries

Intermediate Results Indicators

1. Percentage of new ACCES student loans given to students who qualify for a government subsistence grant 2. Percentage of new ACCES student loans given to students who qualify for a government subsistence grant and score below the SISBEN cutoff points 3. Percentage of new ACCES student loans given to indigenous peoples, afro-Colombians, Roma, and victims of violence 4. Geographic distribution of ACCES loans to new students in zonas apartadas


1. 6.27–8.07 pp increase in the proportion of students from vulnerable families enrolling in higher education 2. 41% increase in participation of the B-20 in higher education 3. More than doubling in the proportion of loans allocated to the poorest students who qualify for maintenance subsidies 4. 40.9 pp increase in the proportion of loans allocated to vulnerable minorities 5. 1.14 pp decrease in the proportion of loans going to students from zonas apartadas

A


Page 43 of 73

ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION

A. TASK TEAM MEMBERS

Name Role

Preparation

Supervision/ICR

Pedro Cerdan-Infantes Task Team Leader(s)

Sandra Ximena Enciso Gaitan Procurement Specialist(s)

Flor Maritza Martinez Camargo Financial Management Specialist

Raul Tolmos Environmental Safeguards Specialist

Javier Botero Alvarez Team Member

Carlos Alberto Molina Prieto Social Safeguards Specialist

Uriel Kejsefman Team Member

Carlos Alberto Roa Cardona Team Member

B. STAFF TIME AND COST

Stage of Project Cycle Staff Time and Cost

No. of staff weeks US$ (including travel and consultant costs)

Preparation

FY13 0 8,042.97

FY14 49.591 222,272.29

FY15 11.785 47,502.69

FY16 8.800 49,236.08

Total 70.18 327,054.03

Supervision/ICR

FY15 8.696 65,902.37

FY16 17.721 89,530.20


Page 44 of 73

FY17 18.802 92,513.75

FY18 16.473 109,365.09

FY19 12.493 65,971.29

Total 74.19 423,282.70


Page 45 of 73

ANNEX 3. PROJECT COST BY COMPONENT

Components Amount at Approval

(US$, millions)

Actual at Project Closing

(US$, millions)

Percentage of Approval

Component 1: Student Loans 424.00 424.00 100.00

Component 2: Institutional Strengthening

12.00 12.00 100.00

Total 436.00 436.00 100.00


Page 46 of 73

ANNEX 4. EFFICIENCY ANALYSIS

1. This analysis follows most of the specifications of the economic analysis developed during the appraisal of the project. It thus seeks to remain comparable to the original analysis. The evaluation is centered on a cost-benefit analysis from the perspective of each major stakeholder that was affected by the project. This type of cost-benefit analysis is rare due to, among other reasons, the difficulty in estimating the social impact of these kinds of investments.7

2. The study includes the direct costs of tuition and the loans as well as features such as the opportunity costs or the incremental contributions that graduate students benefiting from the project will make to the social security system administered by the GoC. Moreover, it incorporates the causal effects of the loan program and the subsidies embedded in the loan rates paid by students. The impact evaluation with causal inference repercussions was conducted independently of these analyses but becomes valuable to evaluate the most direct and likely effect of PACES.

3. Assumptions. Most of the assumptions from the appraisal cost-benefit (ACB) analysis are kept and are updated with the latest data. The main assumptions are as follows:

(a) The working life of students is set at 40 years after graduation or dropping out, regardless of study program length and socioeconomic conditions.

(b) The increase in the wages of students pursuing higher education is the main source of benefits. The records of graduates are obtained from the Ministry of Education Labor Observatory for Education (Observatorio Laboral para la Educación), and a Mincerian approximation is made based on current recommendations,8 using an unconditional quantile regression.9 Furthermore, the mean wages are approximated by the 56th percentile, considering the long tail that the distribution exhibits and previous difference between the observed mean wages and the 50th percentile. This analysis is conducted separately for the different levels of education of ACCES beneficiaries. To obtain the wage for a dropout and for a high school graduate (the opportunity cost counterfactual), we use the Mincer coefficient from the ACB (0.0086), which is the coefficient found in Sanchez and Núñez (2011).10

(c) The starting wages after one year of graduation in 2016 was COP 1.78 million for a college degree, COP 1.13 million for a technological degree, and COP 1.01 million for a technical degree.

(d) These wages are adjusted to the expected inflation (3 percent) and economic growth (3 percent) per year based on the inflation target of the Colombian Central Bank (Banco de la

7 Jimenez, E., and H. A. Patrinos. 2008. Can Cost-Benefit Analysis Guide Education Policy in Developing Countries? World Bank. 8 Heckman, J. J., L. J. Lochner, and P. E. Todd. 2006. “Earnings Functions, Rates of Return and Treatment Effects: The Mincer Equation and Beyond.” Handbook of the Economics of Education 1: 307–458. 9 Firpo, S., N. M. Fortin, and T. Lemieux. 2009. “Unconditional Quantile Regressions.” Econometrica, 77 (3): 953–973. 10 Torres, F. S., and J. N. Méndez. 2003. “A Dynamic Analysis of Human Capital, Female Work-force Participation, Returns to Education and Changes in Household Structure in Urban Colombia, 1976–1998.” Colombian Economic Journal 1: 109.


Page 47 of 73

República) and the International Monetary Fund macroeconomic forecasts, respectively; this becomes the central measure to adjust cohort-specific wages.

(e) These adjusted wages are then projected with the coefficients of the polynomial of experience, obtained in the unconditional quantile regression conducted for each type of degree. To find the income growth for high school graduates, the growth rate of different diplomas is averaged under the assumption that they follow parallel trends, following the original ACB.

(f) Considering the same earnings profile for dropouts and high school graduates is a conservative assumption that sets estimates of returns on the lower bound. In fact, dropouts are expected to have higher incomes than high school graduates, so these simulations are somewhat below their actual expected income.

(g) The time length of the programs is set at 10, 8, and 6 semesters for a BA, technological, and technical degrees, respectively.

(h) Dropout cohort rates according to the System for the Prevention of Dropouts in Higher Education (Sistema para la Prevención de la Deserción de la Educacion Superior, SPADIES) are 44.8 percent for a 10-semester college cohort, 57.6 percent for an 8-semester technological cohort, and 59.4 percent for a 7-semester technical cohort. These rates are adjusted following the results of the impact evaluation of the ACCES student loans11 for students without a subsidy and are further adjusted for those students who received a subsidy, here following the impact evaluation on maintenance subsidies embedded in the ACCES loan.12

(i) Other assumptions considered are the following: (i) the unemployment rate for graduates is set at 5 percent and for dropouts and high school graduates at 10 percent; (ii) the expected loss from default is set at 15 percent for graduates and at 50 percent for student dropouts; (iii) to estimate the returns for the GoC, the key parameter is the current contributions to the social security system (12.5 percent of all declared income); (iv) in the social security system, the GoC recognizes a cost on a per capita basis set by the Ministry of Health (Unidad por Capitación)—this value is used as part of the opportunity cost for the student who graduates and as savings for the GoC; (v) the administrative costs for ICETEX are 0.5 percent for disbursements and 0.5 percent for debt collection, to account for the administrative allocation and collection costs; and (vi) following the ACB, the cost per student for an institution is estimated at 50 percent of the tuition value.

4. Methodology. The basic structure for cost-benefit analyses in the field of education was developed by Becker (1964). That framework is expanded to include other streams of cash in the project and analyze the effects beyond students and entities involved in the provision of the program. The analysis

11 Melguizo, T., F. Sanchez, and T. Velasco. 2016. “Credit for Low-income Students and Access to and Academic Performance in Higher Education in Colombia: A Regression Discontinuity Approach.” World Development 80: 61–77. 12 Lozano-Rojas, Felipe. 2018. Effects of Cash Subsidies Embedded in Student Loans. SSRN. https://ssrn.com/abstract=3208286.

https://ssrn.com/abstract=3208286


Page 48 of 73

considers income streams from social security contributions and subsidy savings from students who graduate but would otherwise be in a subsidy regime were it not for the degree they are attaining.

5. Students’ cost-benefit is determined by the net extra income they make, on top of what they would have made otherwise, minus the cost of attendance and the cost of the loan. Students pursuing higher education delay their labor market entry in exchange for higher wages in the future. The timing of the payments and the total cost of attendance vary, as students pay the cost of their attendance at the time of paying off the loans, and not at the time in which they pursue their studies.

6. ICETEX obtains resources from loan payments that it does and will receive, minus administrative costs and default rates. The third stakeholder considered is the GoC, which grants resources for subsidies. These include the interest rate subsidy and the maintenance subsidies embedded in ACCES student loans, plus the subsidy assigned by the Ministry of Health to the social security subsidized regime. Qualifying students who enjoy the latter subsidy are expected to lose it once they enter their labor market. Hence, the GoC has these savings plus tax collections in graduates’ contributions to the social security system. Dropouts and high school graduates are more likely to stay in the subsidized regime and need state resources.

7. Finally, the cash flows for the HEIs are calculated from the tuition income they receive from students and ICETEX, minus the cost per student. All cash flows are constructed for graduates and dropouts while weighting the final cashflows by the probability of dropping out. For simplicity, students are assumed to drop out in the third semester for the college degree and in the second semester for professional technical and technological higher education institutions (T&T) beneficiaries.

8. Table 4.1 summarizes the set of cash flows while pointing out the different transfers that happen across the different actors. As shown in the last column, when all the actors are considered, most cash flows even out. Consequently, the main source of value of the project is the increased income of students pursuing higher education, whereas the net cost of the project is derived from the education provision cost. Although this economic analysis discusses the rates of return from the perspective of the different stakeholders, the project’s internal rate of return (IRR) results from the sum of all the cash flows.

Table 4.1. Assumptions for cost benefit analysis, by actor

Students/ Graduates

ICETEX GoC HEIs (Total)

− Attendance cost − Attendance cost

− Tuition cost + Tuition cost 0

+ ICETEX

disbursements

− ICETEX

disbursements

0

+ Cash subsidy − Cash subsidy 0

+ Interest rate

subsidies

− Interest rate

subsidies 0

− Interest rate

subsidies

+ Interest rate

subsidies

0

− Education provision

costs

− Education provision

costs

+ Income + Income


Page 49 of 73

Students/ Graduates

ICETEX GoC HEIs (Total)

− SS

contributions

+ SS

contributions 0

− SISBEN subsidies + SISBEN subsidies 0

− ICETEX loan

repay

+ ICETEX loan

repay

0

+ Interest rate

subsidies

−Interest rate

subsidies 0

− Interest rate

subsidies

+ Interest rate

subsidies

0

Source: Authors’ mapping of project actors based on Nakajima (2016).

I. Individual Students Cost-Benefit Scenario

9. Following the ACB, the costs faced by an ACCES loan beneficiary are (a) attendance when the loan program does not cover 100 percent of tuition or other additional expenses; (b) the opportunity costs associated with income forfeited during the study period and beyond, which they would have received in the absence of the loan; (c) the financial cost of the loan, given by the repayment of both principal and interests of the loan. To these three categories, the incremental contributions/subsidies that a beneficiary makes/receives related to the social mobility after graduation are extended. If students graduate, they stop receiving subsidies in the social security system and become contributors instead.

10. The cost and benefit of an individual student is then calculated by Equation 1:

(1) 𝑆𝑡𝑢𝑑𝑒𝑛𝑡 𝐵𝑒𝑛𝑒𝑓𝑖𝑡𝑖𝑡 =

∑ (𝑊𝑖𝑡

𝑖=−𝑛

− 𝑊𝑏𝑡) − (𝐴𝑡𝑡𝑒𝑛𝑑𝑎𝑛𝑐𝑒 𝑐𝑜𝑠𝑡𝑖𝑡 + 𝑇𝑢𝑖𝑡𝑖𝑜𝑛 𝑐𝑜𝑠𝑡𝑖𝑡 + 𝐶𝑎𝑠ℎ 𝑆𝑢𝑏𝑠𝑖𝑡 + 𝐿𝑜𝑎𝑛 𝑐𝑜𝑠𝑡𝑖𝑡

+ Δ 𝑆𝑆𝐶𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠𝑖𝑡) where 𝒊 is the type of student in question (bachelor’s, technological, or technical), the length of the program (5, 4, or 3 years, respectively), the individual wage (𝑾𝒊𝒕), the average disbursement (𝑻𝒖𝒊𝒕𝒊𝒐𝒏 𝒄𝒐𝒔𝒕𝒊𝒕) plus an additional cost of attendance calculated as 30 percent of the average disbursement, the fact of receiving or not maintenance subsidies, and the contribution and net of subsidies lost (𝚫 𝑺𝑺𝑪𝒐𝒏𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔𝒊𝒕). 𝑾𝒃𝒕 is the instantaneous wage of a high school graduate, which represents the most important opportunity cost. (𝑾𝒊𝒕 − 𝑾𝒃𝒕) represents the incremental change in income (𝚫 𝑰𝒏𝒄𝒐𝒎𝒆) due to attaining graduation, and thus 𝑾𝒊𝒕 is found by probabilistically weighting the graduate cash flow to the probability of graduation and the dropout cash flow to the probability of dropping out.

11. Relative to table 4.1, other elements of the equation should be thought as transfers from and to other actors, and their final sign would depend on the direction of the cash flow. For instance, 𝑳𝒐𝒂𝒏 𝒄𝒐𝒔𝒕𝒊𝒕 includes the original disbursements and the repayments by the student.


Page 50 of 73

Inputs Update

12. Table 4.2 summarizes the updates on the representative students that undertake a technical degree. Wages change according to the latest information as well as the periodic disbursements and thus, the total debts estimated at the time of repayment starts for individuals who graduated and funded their studies completely with student loans. Length of programs and labor activity remains the same. Cash subsidies are updated to the information disclosed by the corresponding entity on a yearly basis, as well as the interest rate charged to students in each period.

Table 4.2. Technical Degrees Input

ACB Current Update

ACB Current Update

With high school 538,365 (as of 2013)

752,757 (as of 2016)

With additional technical degree

1,003,502 (as of 2013)

1,013,489 (as of 2016)

Length (semesters) 6 6 Length (years) 3 3

Disbursement − with subsidy students (2014-2)

1,768,500 (as of 2013)

2,129,075 (as of 2017)

Total debt at repayment start − with subsidy (2014-2)

10,611,000 (as of 2013)

11,896,651 (2014 cohort

at repayment)

Disbursement − without subsidy (2014-2)

1,800,883 (as of 2013)

2,129,075 (as of 2017)

Total debt at repayment start − without subsidy (2014-2)

10,805,298 (as of 2013)

14,375,398 (2014 cohort

at repayment)

Time span 40 years 40 years Cash subsidy 669,444 (2013)

831,398 (2018)

Inflation (rate for subsidized students)

2.44% Yearly inflation and 3% from

2018 onwards

Inflation + addition (students without subsidy)

6.44% IPC + 4%

12% from 2018 onward

IPC + 9%

Source: Authors’ calculations based on ICETEX, SPADIES, Observatorio Laboral de la Educacion, Sánchez and Núñez (2003), and the ACB Analysis of the ACCES II Program. Note: Values in nominal COP unless specified otherwise.

13. Table 4.3 presents the inputs updated for students pursuing a technological degree.

Table 4.3. Technological Degrees Inputs

ACB Current Update

ACB Current Update


752,757 (as of 2016)

With additional technological degree

1,003,502 (as of 2013)

1,132,527 (as of 2016)

Length (semesters) 8 semesters 8 semesters Length (years) 4 years 4 years

Disbursement − with subsidy students

1,768,500 (as of 2013)

2,058,277 (as of 2017)

Total debt at repayment start − with subsidy

10,611,000 (as of 2013)

16,710,505 (2014

cohort)

Disbursement − without subsidy

1,800,883 (as of 2013)

2,058,277 (as of 2017)

Total debt at repayment start − without subsidy

10,805,298 (as of 2013)

20,750,425 (2014

cohort)

Time span 40 years 40 years Cash subsidy 669,444 831,398


Page 51 of 73

ACB Current Update

ACB Current Update

(2013) (2018)


2.44% Yearly inflation and 3% from 2018 onwards


6.44% IPC + 4%

12% from 2018

onwards IPC + 9%

Source: Authors’ calculations based on ICETEX, SPADIES, OLE, Sánchez and Núñez (2003), and the ACB Analysis of the ACCES II Program. Note: Values in nominal COP unless specified otherwise.

14. Finally, the update of inputs for the university students is presented in table 4.4.

Table 4.4. University Degrees Inputs

ACB Current Update

ACB Current Update


752,757 (as of 2016)

With bachelor’s university degree

1,620,220 (as of 2013)

1,780,638 (as of 2013)

Length (semesters) 10 semesters 10 semesters Length (years) 5 years 5 years

Disbursement − with subsidy students (2014-2)

1,941,485 (as of 2013)

3,431,880 (as of 2017)

Total debt at repayment start – with subsidy (2014-2)

19,144,845 (as of 2013)

36,533,538 (2014

cohort)

Semester disbursement − without subsidy (2014-2)

1,418,400 (as of 2013)

3,431,880 (as of 2017)

Total debt at repayment start – without subsidy (2014-2)

14,183,995 (as of 2013)

46,503,433 (2014

cohort)

Time span 40 years 40 years Cash subsidy 669,444 (2013)

831,398 (2018)


2.44% Yearly inflation and 3% from 2018 onwards


6.44% IPC + 4%

12% from 2018

onwards IPC + 9%

Source: Authors’ calculations based on ICETEX, SPADIES, OLE, Sánchez and Núñez (2003), and the ACB Analysis of the ACCES II Program. Note: Values in nominal COP unless specified otherwise.

Earning Profiles

15. The earnings profile of higher education graduates is presented in figure 4.1. As the analysis is nominal, wages continue to increase at the inflation rate and at the economic growth rate after the profile from OLE data is exhausted (after 12 years of experience). The graph also shows the nominal earnings profiles used in the estimation, once the experience profile from the OLE is exhausted, as salaries are allowed to grow at the expected inflation rate (3 percent).


Page 52 of 73

Figure 4.1. Yearly Earnings Profile: Higher Education Graduates

Source: Author’s calculations using OLE

Individual Students Rates of Return

16. The original ACB referred to three different types of students per the degree level: (a) Strata 1 and 2 with subsidy, (b) Strata 1 and 2 without subsidy, and (c) other strata without a subsidy. It is simplified to two representative students, with and without subsidy, given that the project stopped funding students from upper strata households with the introduction of Tú Eliges that targeted vulnerable students, paying either 0 or 10 percent of their balance during the semester of studies and diverted the rest to other loan lines. Accordingly, less than 7 percent of beneficiaries in 2014-2 and 2015-1 were from the richest strata.

17. ACCES II was also affected by the removal of the interest rate subsidy based on strata and the unification of the SISBEN classification of subsidies beneficiaries. Before Tú Eliges, a maintenance subsidy was awarded based on SISBEN scores, and a separate interest rate subsidy was recognized by the GoC to ICETEX for the funding of students belonging to the first three strata according to their residence utility bills. After the introduction of Tú Eliges, both maintenance subsidies and interest rate subsidies were awarded on a SISBEN threshold basis. For simplicity, calculations are classified into two types of students regardless of their actual levels of interest rate. ‘No subsidies’ students refer to students paying full cost interest rates (IPC + 9 percent according to Tú Eliges changes) and ‘with subsidies’ to those whose student profile includes maintenance subsidies and the interest rates subsidies embedded in the student loan. Representative students have the following IRRs:

18. Relative to the ACB analysis, the current estimation presents a substantial reduction. The ACB found IRRs of 24 percent for technical students, 19 percent for technological students, and 15 percent for college students. The main reasons for this reduction are (a) the inclusion of further contributions of the

BA College AD Technological AD Technical

No Subsidies 9.39% 5.55% 5.26%

With Subsidies 11.98% 6.79% 6.13%

IRR Student


Page 53 of 73

students which reduce their income (while the original ACB did not consider payments to social security, this analysis does include them to allow for calculating a rate of return for the GoC) and (b) the update of data shows a severe increase in the cost of attendance, which is explained, to some extent, by the fact that ICETEX started financing 100 percent of tuition with the introduction of Tú Eliges.

19. The following table is calculated without discounting the contributions made by the students to the social security system. The IRR for college graduates with subsidies then increases to 14 percent, close to the ACB level. However, other levels of education are affected disproportionately. Despite this comparison, investing in higher education remains a very valuable and profitable investment for Colombian students as the IRR is considerably positive.

II. ICETEX’s Perspective: Individual Student Loan Cost-Benefit Scenario

20. From the ICETEX perspective, the cost of the project is related to the disbursements of student loans to HEIs. The return on the investment is given by two separate flows: (a) student repayments and (b) interest rate subsidies that the GoC recognizes for eligible students. Traditionally the Government has recognized a differential, matching the cost per student to ICETEX.

21. The following is the ICETEX cash flow equation:

(2) 𝐼𝑐𝑒𝑡𝑒𝑥 𝐶𝐹𝑖𝑡 =

∑ 𝑆𝑡𝑢𝑑𝑒𝑛𝑡 𝑅𝑒𝑝𝑎𝑦𝑚𝑒𝑛𝑡𝑠𝑖𝑡(1 − 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝐿𝑜𝑠𝑠𝑖) + 𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑆𝑢𝑏𝑠𝑖𝑑𝑖𝑒𝑠𝑖𝑡 − 𝐿𝑖𝑡 − 𝐴𝑖𝑡

1+2𝑛

𝑖=−𝑛

22. As in Equation 1, Equation 2 is the outcome of a weighted average of separate cash flows representing graduate and dropout individuals. ICETEX cash flow in period 𝒕 for student type 𝒊 is equal to its incomes from the students’ repayments, net of any default, as measured by the expected loss of the student type, and the transfers for interest subsidies minus its cash uses in loan disbursements (𝑳𝒊𝒕) and in its administration costs (𝑨𝒊𝒕). This specification allows to construct a vector of investments and returns that lead to an IRR per student type. The repayment period is established for twice the length of the studies plus a grace period before the start of the repayment. The following table presents the IRR that ICETEX receives for each of the individualized student type loan operations:


No Subsidies 11.42% 7.71% 7.64%

With Subsidies 14.41% 9.41% 9.22%

IRR Student - No SS Contributions nor Income Tax witholding


No Subsidies 9.39% 8.11% 7.01%

With Subsidies 14.98% 14.37% 13.38%

IRR ICETEX - Individual Students


Page 54 of 73

23. Relative to the ACB, the IRRs found in this analysis are higher. The ACB finds negative rates for ICETEX when investing in T&T students and a 1 percent IRR from loans for college students. However, it is not clear how the calculations were made for the interest rate subsidies not for the payments net of nonperforming loans in default. The calculation of the interest rate subsidy requires the disclosure of the interest rate at which the original loans are allocated as ICETEX charges the student one part and the remaining is charged to the GoC in the subsidized loans. The reference interest rate in 2014 was 15.77 percent EAR according to ICETEX records, from which students paid the last available consumer price index and the Government recognized the difference. When analyzing the loans without the interest rate subsidy, for the subsidized students one sees that the returns for ICETEX drop substantially. The most likely difference, it seems, is the specification assumption used for the interest rate subsidy; however, the ACB does not disclose how it is introduced. Another possible explanation is the reduction in the dropout rates following the introduction of the outcomes of the impact evaluations conducted over the ACCES program. By reducing the dropout rates, ICETEX improves collection, which renders a higher rate of return for ICETEX.

III. Cost-Benefit for the GoC

24. The third stakeholder considered in this analysis is the GoC. The GoC makes yearly investments supporting student loans and its beneficiaries to foster academic success of beneficiaries and to support the independence and sustainability of ICETEX. The GoC makes these investments through the Ministry of Education and operationalizes it through ICETEX itself.

25. One of the channels through which the society benefits from investing in education is the higher tax base created by the income increase that students register after graduation. Accordingly, students who successfully graduate make social contribution payments of 12.5 percent of their declared income as registered in the OLE information system. Finally, students who start making contributions are expected to leave the subsidized regime in social security. This would represent significant savings for the GoC, given that through the Ministry of Health, it recognizes the beneficiaries of the subsidized system on a per capita basis that can be easily accommodate in the cash flow.

26. The following equation summarizes the cash flows that the GoC experiences owing to the ACCES loan program:

(3) 𝐺𝑜𝐶 𝐶𝐹𝑖𝑡 =

∑ 𝑆𝑆 𝑆𝑢𝑏𝑠 𝑆𝑎𝑣𝑖𝑛𝑔𝑠𝑖𝑡 + Δ𝑆𝑆 𝐶𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠𝑖𝑡 − 𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑆𝑢𝑏𝑠𝑖𝑡 − 𝐶𝑎𝑠ℎ 𝑆𝑢𝑏𝑠𝑖𝑡

𝑖=−𝑛

,

where 𝒊 identifies the student type in period 𝒕. The GoC experiences savings from students who will not need the social security subsidies and collects taxes from social security contributions. The investment the


No Subsidies 9.39% 8.11% 7.01%

With Subsidies 11.98% 0.97% 0.32%

IRR ICETEX - Individual Students - No Interest Rate Subsidy


Page 55 of 73

GoC makes consist of subsidies for student loans, the interest rate subsidy, and the maintenance subsidies.

27. The following table presents the IRR for those students who receive investments from the GoC. In this setting, the IRR cannot be calculated for students without subsidies as in their case there are no investments from the Government.

28. The highest rate of return is for T&T students. The longer-lasting balances and higher costs in college degrees decrease the return of college students relative to T&T students for the GoC, who pay faster and do not require longer-term interest rate subsidies. The ACB did not include an analysis recognizing the cash flows for the GoC. However, in the analysis, the ACB includes education spillovers and a decrease in crime, which are not included here as accurate data are not available.

IV. Aggregate Cost-Benefit Analysis for the Society as a Whole

Total Number of Students Funded

29. To find the aggregate cash flows, first the aggregate cash flows of the individual student types are estimated according to the number of new students as reported by ICETEX. During the length of the project, ACCES II Phase II ICETEX funded 111,615 students exclusively in the lines affected by the loan project with the World Bank. Figure 4.2 presents how those students are distributed among the student types according to their level of studies or if they are eligible for subsidies or not. These individuals are beneficiaries of ACCES loans before 2015-2 and from that semester on, are beneficiaries of ACCES - Tú Eliges 0 percent and ACCES - Tú Eliges 10 percent.


With Subsidies 8.57% 11.05% 14.55%

TIR ICETEX - Individual Students - No Interest Rate Subsidy


Page 56 of 73

Figure 4.2. New Students Loans in ACCES

Source: ICETEX

30. From the total, 95 percent of the beneficiaries pursue a university degree, and 36 percent start their loans as beneficiaries without subsidies. It has been found that some students change their status as they review their SISBEN or as they realize that they are eligible for the subsidies and migrate from one status to the other. For simplicity, students follow the status that they were at the inception of the semester cohort, regardless of any movement in their subsidy status.

31. Figure 4.2 shows an important reduction in the number of students funded per semester after the introduction of Tú Eliges in 2015-2, with the exception of 2016-1. It is assumed that this is the result of the cap established in the number of students receiving interest and maintenance subsidies introduced by the Ministry of Education to prevent the depletion of sectorial resources from the investment budget allocated to education by the GoC. The cap established that only 20,000 students would receive subsidies, and this limited the scope of growth for the ACCES - Tú Eliges loan programs which focus on favorable conditions for the most vulnerable students.

Extending the Cohorts

32. With the number of beneficiaries per cohort, representative students are estimated accordingly. Cash flows increase for each cohort using an accumulated inflation from 2014-2. Using the projection for each of stakeholder, the following rates of return are obtained:

33. These rates are obtained by extending the aggregated cash flows and including the cash flows for HEI as well. IRRs for HEIs are not found as they take a net benefit from each student by discounting from

BA College AD Technological AD Technical Total

w/ Subs 13.80% 9.17% 9.08% 13.43%

w/o Subs 12.95% 8.71% 8.48% 12.67%

Total 13.25% 8.94% 8.72% 12.96%

Total

Aggregate IRR - Students + Icetex + GoC + HEIs


Page 57 of 73

the tuition their total costs; it is assumed, following the previous ACB, that their net benefit is 50 percent of the tuition. The following figures present the construction of the aggregate cash flows differentiating the net flows that pertain to each stakeholder.

Figure 4.3. Total Cash Flow—with Subsidy

Source: Own Calculations

Figure 4.4. Total Cash Flow—without Subsidy

Source: Own Calculations

34. The total IRR of the project adding all the stakeholders’ cash flows is 12.96 percent, which remains high and substantial. Although the aggregate benefit found here is lower than the one in the ACB (where the IRR was 18 percent), this analysis does not take into account some of the societal returns that are considered in the ACB such as the decrease in crime and the education spillovers.


Page 58 of 73

Conclusions

35. The individual return to higher education of the representative student in Colombia remains high, with IRRs ranging from 5.3 percent to 14.5 percent depending on the educational level pursued and the amount of the subsidy that the GoC invests in the student’s education. However, compared to the previous economic analysis, these values are different mainly due to the increase in the cost of attendance experienced by Colombian students during the project implementation and by the latest data in the inputs. Still, this increase in the average disbursement is aligned with the project goals, as it reduces the cost of attendance during the period of studies, thus allowing the poorest and those with limited liquidity to access education independently of their financial situation during the period of studies.

36. ICETEX’s IRRs, on the other hand, improved substantially, moving from barely positive in the ACB to a range between 7 percent and 14 percent, depending on the type of student. This difference is attributed to the way that interest rates subsidies are calculated, with the methodology applying the actual accrual of interests that the GoC recognizes and to the decrease in dropout rates following the impact evaluations conducted on the ACCES program (Melguizo, Sanchez, and Velasco 2016; Lozano-Rojas 2018). Smaller dropout rates among beneficiaries render better degrees of collection and thus better rates of return for ICETEX.

37. Total returns to investing in higher education remain substantially high at 13 percent. This rate is smaller relative to the Project Appraisal Document’s economic analysis. This change is explained partially by some of the externalities of investing in higher education that are not considered here, the decrease in crime, and the spillovers of education on other workers. Relative to the IRR found in the ACB, when these two assumptions are dropped, the IRR becomes 15 percent, within the range of this analysis’s findings. This economic analysis then shows that there has not been a substantial change in the returns to higher education for society as a whole.


Page 59 of 73

ANNEX 5. BORROWER COMMENTS

This is a summary of the Borrower’s Final Report of the ACCES 2 SOP-Phase 2 Project (Informe de Cierre Proyecto ACCES 2 SOP-Fase 2. BIRF 8354-CO), written in May 2018 and translated into English and redacted by the World Bank. Any mistakes are the sole responsibility of the Bank.

EXECUTIVE SUMMARY OF ACCES II PHASE 2

ACCES and Tertiary Education in Colombia

1. The ACCES Project introduced the country’s first socially oriented long-term subsidized educational line of credit. During its first stage (2003-2007), it benefited more than 124,000 new students with over 95 percent of loans being allocated among the lowest strata of the population.

2. The ACCES II project was implemented and divided into two phases of execution. The first phase was executed between 2008 and 2012, financed through loan BIRF 7515-CO of US$300 million, benefitted 170,000 students with new loans and more than 930,000 re-registrations. The second phase, presented in this document, was executed on December 22, 2014, upon the signing of the contract of loan BIRF 8354-CO of US$200 million, structured to keep boosting demand while seeking to reinforce other aspects like quality and equity, including the institutional strengthening of ICETEX (where the counterpart contribution of the entity stood at US$236 million for an aggregate project amount of US$436 million).

3. By 2014, before the beginning of the second phase of ACCES II, the country’s rate of coverage for undergraduate studies was 47.8 percent. Nevertheless, the country’s challenge in tertiary education remained huge since Colombia’s rate of coverage is far from the one of the OECD countries (75 percent). For this reason, the PND 2014–2018 ‘Todos por un nuevo país’ (Everyone for a new country) set a goal of coverage in tertiary education of 57 percent by 2018, along with a focus on the quality of the system (the latter to reduce dropout rates and improve graduates’ employability). In such a context, ACCES II Phase 2 was essential in contributing to finance tuition fees for the country’s least favored population.

Purpose of ACCES II Phase 2

4. The project was intended to increase students’ registration and graduation rates, as well as equity in tertiary education by (a) increasing the number of ACCES student loans with a focus on students from households in unfavorable socioeconomic situation, (b) promoting student loans in high-quality institutions and programs, and (c) enhancing the institutional capacity of ICETEX.

ACCES II Phase 2: Implementation and Results

Financial Execution of the Project

5. The execution of the US$200 million loan began in March 2015, while the disbursement stage finished in 2017 (see Table 5.1). Even though its closing was planned for 2019, with the introduction of the education loan model Tú Eliges (You choose) which finances 100 percent of tuition fees, the average value of transfers rose about 100 percent, which required a faster execution of funds.


Page 60 of 73

Table 5.1. Execution of World Bank Resources (ACCES II Phase 2)

Source: Planning Advisory Office - ICETEX.

Accomplishment of the Project Development Objectives (PDOs)

6. Number of students enrolled in tertiary education (undergraduate studies). Increased access to tertiary education is the greatest achievement in terms of the ACCES II Phase 2 PDOs. The number of students enrolled in programs of undergraduate studies rose from 2,080,440 in 2014 to 2,280,327 in 2017, which accounted for an increase of 9.6 percent in enrollment for the period with an annual average growth of 3.1 percent. This is a direct result of the emphasis on coverage as a government policy aiming to achieve the goal of 57 percent in 2018, which, by the end of 2017, already stood at 52.8 percent.

7. Equity measured as the percentage of new students graduated in tertiary education coming from families from disadvantaged backgrounds. Graduation equity was measured as the percentage of students from families with an income of two monthly minimum wages graduated in tertiary education with respect to the total of graduated students in undergraduate programs. The project featured an average accomplishment level of 101.8 percent, as compared to the annual goals set for 2014–2016.

8. Increase of equity in tertiary education: percentage of new students registered in tertiary education with an income lower than two minimum wages. The percentage of registered students coming from families with income lower than two minimum wages grew almost 1 percentile, going from 59.92 percent as compared to the estimation of 60.82 percent, versus the growth goal of 2 percent in the period. Given this estimation, the project reported a level of accomplishment of 97.3 percent as of 2017.

9. Nevertheless, progress is evidenced by checking other equity indicators concerning access to tertiary education. In 2016, 78.2 percent of students pursuing the first course of SENA were reported to come from families with income of less than two minimum wages. Moreover, from 2010 to 2016, population at level 1 of SISBEN13 (the most vulnerable) increased by 11.78 and student enrollment rates of SISBEN beneficiaries went from 47.1 percent to 60.4 percent. Furthermore, student enrollment in tertiary education from the quintile 1 of income increased by 4.5 percent between 2008 and 2017 and quintile 2, by 6.6 percent (see Figure 5.1).14 Finally, net coverage rose by 6.9 percent for quintile 1 of income between 2008 and 2017, and for quintile 2, it rose by 7.3 percent (see Figure 5.2).

13 SISBEN is the System of Identification of Potential Beneficiaries of Social Programs. Source: https://www.sisben.gov.co/sisben/. 14 In accordance with an analysis of the Deputy Directorate of Education of the National Planning Department (Departamento Nacional de Planeación, DNP), based on figures published by the National Directory of Statistics (Dirección Nacional de Estadística, DANE) (Survey on Quality of Life or Encuesta de Calidad de Vida - ECV).

Fecha Solicitud No. TRMValor

desembolsado US $

Valor desembolsado

COP

06/03/2015 1 2565,61 27.086.969,20 69.494.599.049,21

19/06/2015 2 2548,2 32.451.467,23 82.692.828.795,49

22/02/2016 3 3163,25 15.764.465,48 49.866.945.429,61

15/06/2016 4 2972,97 14.973.369,19 44.515.377.400,79

23/08/2016 5 2883,89 15.137.835,74 43.655.853.112,23

09/09/2016 6 2846,13 15.487.362,67 44.079.047.515,97

04/01/2017 7 2981,06 39.982.811,53 119.191.160.139,62 25/01/2017 8 2932,01 39.115.719,01 114.687.679.294,51

200.000.000,05 568.183.490.737,43 Total

https://www.sisben.gov.co/sisben/


Page 61 of 73

Figure 5.1. Distribution of Enrolled Students per Quintile of Income

Figure 5.2. Net Coverage Rate per Quintile of Income

Source: DNP. Calculations based on the ECV - DANE. Source: DNP. Calculations based on the ECV - DANE.

10. Number of students graduated in tertiary education. The project estimated to go from 289,000 students graduated in 2014 to 375,000 in 2017. Nonetheless, for this indicator a total of 334,440 graduated students were observed in 2016 and 365,160 graduated students in 2017 for an average level of accomplishment of 96.4 percent with respect to the annual goals. The annual average rate of graduation, although below the project goal, has evolved positively, going from 4.7 percent between 2001 and 2010 to about 6.7 percent between 2012 and 2017. With regard to desertion, SPADIES data show that college desertion per period has been improving, going from 12.9 percent in 2010 to 9 percent in 2016.

Performance of ICETEX Education Loans

11. ACCES II Phase 2 included improvements aimed at favoring permanence, equity, and quality education. A paramount strategy was to extend the scope to offer more line of credit options for students under vulnerable conditions. This strategy sought to reach to people with the lowest SISBEN scores and special populations (ethnic groups and victims of violence). Additionally, differential conditions were offered in more than 10 departments with high indexes of unsatisfied basic needs (UBN) (considered as remote areas). Nonetheless, obstacles such as low demand and low presence of tertiary education institutions in those areas have hindered an optimal result in this domain.

12. Number of new and renewed loans. During the period of analysis, a total of 148,460 new loans were granted in the ACCES lines at a national level (figure 5.3) which also had an impact on the number of renewed loans during the period, which amounted to 806,843 (figure 5.4). The difficulty in reaching the proposed goal was mainly related to the implementation of the program Tú Eliges, which integrated financing of 100 percent of tuition to reduce desertion and give students the opportunity of having access to high-quality tertiary programs and institutions. This increased the average payment values per student, affecting the number of placements.

5.8% 6.5% 7.3% 9.1% 10.3%10.4% 12.2% 13.5% 15.0% 17.0%21.0% 21.2% 20.1%

22.9% 21.6%

31.8% 28.4% 30.7%27.8% 28.0%

31.0% 31.6% 28.4% 25.2% 23.1%

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

2008 2011 2013 2015 2017

quintil 1 quintil 2 quintil 3 quintil 4 quintil 5

8.2% 7.4% 8.7%14.3% 15.1%

12.4% 14.2% 16.5%19.1% 19.7%

23.4% 24.1% 22.8%28.4% 26.7%

35.5% 33.9% 35.6% 37.8%42.2%

47.2%

59.5% 60.2%54.6%

58.3%

2008 2011 2013 2015 2017quintil 1 quintil 2 quintil 3

quintil 4 quintil 5


Page 62 of 73

Figure 5.3. New Loans Granted in 2014–2017 Figure 5.4. Number of Renewed ACCES Loans in 2014–2017

Source: Planning Advisory Office - ICETEX. Source: Planning Advisory Office – ICETEX.

13. Department distribution. The Capital District obtained the highest number of beneficiaries with 24,733 (16.7 percent), followed by Valle del Cauca with 11,343 (7.64 percent), Atlántico with 10,334 (6.96 percent), and Antioquia with 9,671 (6.45 percent). These five departments are focused on about 40 percent of the ACCES II Phase 2 loans for 2014–2017, while departments with the lowest percentage of participation are characterized by high UBN conditions, including Chocó (UBN of 79.19 percent), followed by Guainía (UBN of 60.62 percent) and Amazonas (UBN of 44.41 percent). ICETEX is aware of this problem, and for this reason, during ACCES II Phase 2, granting of the loan was based on a rating model where for each loan committee, requests with a family core corresponding to a remote area15 were allocated first.

14. Level of education. College education had a higher number of new beneficiaries with a total of 132,377 while technical and technological professional education totaled 16,084 beneficiaries. It is worth noting that the value of loans is higher for college education where the highest participation is found.

15. Social and economic strata. Given the scoping policy based on Tú Eliges, beneficiaries from Strata 1, 2, and 3 are predominant. However, there may be some cases of special populations (ethnic groups and victims of violence), as well as beneficiaries belonging to partnerships coming from other strata. Stratum 2 had the highest average participation during 2014–2017 with 43.9 percent, followed by stratum 1 with 37.7 percent in average, which represented altogether 81.6 percent of ACCES II Phase 2 loans.

16. Gender. ACCES II Phase 2 shows a higher participation of female population, with the total beneficiaries of 84,349 (56.8 percent on average), compared to male population, which represented 64,067 (43.2 percent on average). A similar trend is observed in the sector since, according to tertiary education statistics (MEN 2015) for 2013–2015, the average participation of women was 52 percent.

17. Special population. In ACCES II Phase 2, the special population went from 9.0 percent in 2012 to 24.6 percent in 2017. This was possible because of the implementation and follow-up of safeguarding plans, which involved fulfillment of commitments toward differential attention. The group that represented a higher number of new beneficiaries during 2014–2017 was the indigenous population (46.7 percent), followed by the Afro-Colombians (27.0 percent) and the victims of violence (26.3 percent).

15 The remote areas determined since the design of the Project are as follows: Amazonas, Arauca, Caquetá, Chocó, Guainía, Guaviare, San Andrés, Putumayo, Vaupés and Vichada.

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

2014 2015 2016 2017

Observado 42,442 37,019 37,440 31,559

Meta 38,000 40,000 41,000 43,000

5,000

55,000

105,000

155,000

205,000

2014 2015 2016 2017

Observado 188,850 205,394 209,123 203,476

Meta 190,000 200,000 207,000 212,000


Page 63 of 73

Benefits of ACCES II Phase 2

1.1. Access (Increase of Student Enrollment)

18. Tú Eliges. In 2015-II, ICETEX introduced the loan model called Tú Eliges that involved changes to access requirements, payment terms, and financing percentages, offering seven new payment modalities, depending on the loan percentage which can be assumed by the student while pursuing studies: 0 percent, 10 percent, 25 percent, 30 percent, 40 percent, 60 percent, and 100 percent. The long-term lines, ACCES-Tú Eliges (0 percent, 10 percent, and 25 percent), are co-financed by the World Bank and are exclusively addressed to the population at a low socioeconomic level, in which, without leaving aside the academic merit, better payment conditions are offered to socioeconomic strata 1, 2 and 3.

19. Unlike ACCES-Traditional, whose lines covered up to 75 percent of the cost of tuition for students from strata 1 and 2 and up to 50 percent of tuition for those from strata 3, 4, 5 and 6, and existence of a co-debtor was required in all cases, the lines ACCES-Tú Eliges only address the population from strata 1, 2 and 3; cover tuition completely (100 percent); keep the benefits of interest rate subsidies, support, and partial loan forgiveness upon graduation, and as a significant innovation, they include the possibility of applying to a Fund of Guarantees which replaces the co-debtor for applicants not having this support. This guideline has, on the one hand, benefited persons under vulnerable conditions and, on the other hand, guaranteed access to institutions and programs of a better quality (which are more expensive).

20. The impact assessments of the previous phases of ACCES showed a relation between desertion and the economic situation of beneficiaries. Based on this, ICETEX conducted a study on its target population establishing market segmentations to target the poorest population with the best academic performance. In this manner, a group of 39,566 young people was. Additionally, the actions oriented toward improving performance of students during their cycle of education were strengthened. The implementation of Plans of Affirmative Actions (which involve academic, psychological, economic, and other interventions), required through the agreements with the tertiary education institutions, is pending.

21. Additionally, within the framework of Tú Eliges, the line of credit, Protección Constitucional (Constitutional Protection), was created, with a preferential interest rate for undergraduates (which also finances 100 percent of tuition), addressed to ethnic groups, victims of violence, reintegrated armed-conflict-affected people, handicapped persons, and members of Red Unidos (which targets population living in extreme poverty). This line has had priority in allocation. During development of ACCES II Phase 2, 100 percent of the demand was satisfied, and 3,413 people benefitted from 2015 to 2017.

Performance Lines ACCES-Tú Eliges

22. Lines ACCES-Tú Eliges 0 percent, 10 percent, and 25 percent represented 71,679 new loans between 2015 and 2017. The line Tú Eliges 25 percent had the greater degree of acceptance, with an average participation of 54.2 percent over the total of ACCES between 2016 and 2017, followed by the line Tú Eliges 0 percent, with an average participation of 22 percent for the same period.

23. The new policy of education oriented toward the improvement of permanence and promotion of quality, led to an increase in the average values of payment higher than the estimations. As evidenced by DANE’s Tertiary Education Cost Index (see figure 5.5), there was an increase in the costs of tertiary


Page 64 of 73

education between 2014 and 2017 which had an impact on the budget of the entity. ACCES II Phase 2 planed with an average value of disbursement per student of COP 2.5 million biannually (which did not correspond to 100 percent of tuition costs), but the actual average value of money order went from COP 3.64 million in 2015-2 to COP 4.44 million in 2017-2 (see figure 5.6). This allows to conclude that the increase in the average value of money order per student, along with the growth of tertiary education costs, decreased the capacity to finance the expected coverage year to year and affected the placement goals negotiated with the World Bank.

Figure 5.5. Tertiary Education Costs Index (annual variation)

Figure 5.6. Average Value of Money Order in Undergraduate Studies 2015-2 to 2017-2 (COP, million)

Source: DANE. Source: Planning Advisory Office – ICETEX.

1.2. Permanence and Graduation (Decrease of Desertion and Increase in Graduation)

24. Impact of ICETEX loan on desertion. ICETEX developed complementary strategies to follow up beneficiaries on their way to graduation. These strategies include (a) granting of economic support and rate subsidies, (b) partial forgiveness of interests upon graduation, and (c) implementation of Affirmative Actions.16 The latter strategy is worth noting since it required institutions to be co-responsible in decreasing the desertion of students financed with ICETEX loans. Altogether, these mechanisms have proven to be important measures to improve permanence (see Figure 5.7).

Figure 5.1. Desertion per Cohort with the ICETEX Education Loan

Source: SPADIES -MEN 2017.

16 ICETEX has 209 agreements with tertiary education institutions including the fourth clause which requires them to provide differentiated attention to ICETEX beneficiaries coming from special populations and to establish mechanisms for decrease of desertion of the beneficiaries of the education loan.

3.82% 3.88% 3.84%

5.37%

6.46% 6.21%

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

2012 2013 2014 2015 2016 2017

3.64 4.01 4.05

4.664.44

5.99

0

1

2

3

4

5

6

7

2015-2 2061-1 2016-2 2017-1 2017-2 2018-1

Mill

on

es d

e p

eso

s$


Page 65 of 73

25. Indicator of desertion and associated late payment. The desertion and associated Late Payment Index (IDM for its Spanish acronym) was created in 2009 with the purpose of establishing the percentage of funds provided by tertiary educations institutions to the Education Loan Sustainability Fund (Fondo de Sostenibilidad del Crédito Educativo, FSCE).17 For the assessment conducted in 2017, it can be observed that the IDM for 2017 has been the lowest since 2009 (see table 5.2).

Table 5.2. Comparative Table of the 2009–2017 Desertion and Associated Late Payment Index

Source: Planning Advisory Office - ICETEX.

26. Schemes of forgiveness and subsidies of interest rate and support. Subsidies associated with the education loan and partial loan forgiveness upon graduation have proven to be essential instruments of public policy to promote access, permanence, and graduation of students. For this reason, transfers of the nation have been an essential resource for the operation of the ACCES II Phase 2 Project (in 2015, this represented 38.7 percent of the sources of resources for the education loans of ICETEX).

27. Between 2015 and 2017, resources were invested in subsidies at the interest rate of COP 1.22 billion, benefiting on average 291,000 students in the active portfolio. The amount invested in subsidies at the interest rate during ACCES II Phase 2 is 43.5 percent higher than the amount invested in the 5 previous years. As for subsidy of support, during ACCES II Phase 2, COP 322 billion was invested, benefiting on average 83,000 students during this period. The impact assessment made by the World Bank for closing of ACCES II Phase 218 concluded that the subsidy of support granted by ICETEX reduced the probability of desertion in tertiary education by 23 percent and increased the probability of graduation by 12.8 percent, showing the relevance of this type of public investment in human capital.

1.3. Quality

28. ACCES II Phase 2 began a strategic path aimed to improve the quality of tertiary education, stimulating loans requested for the accredited tertiary educations institutions and programs. This quality policy, implemented as an incentive, by assigning a higher score in the loan-granting model for the programs and institutions accredited from ACCES II Phase 2, has had a double effect: (a) it has allowed

17 The FSCE was created to mitigate and cover the risk of loans of the loan lines as a result of desertion during term time to keep sustainability of financing and access to tertiary education. The percentages of contribution of tertiary education institutions are defined according to the level of desertion and associated late payment. ICETEX contributes to the FSCE in an equal proportion, the percentage determined for each institution. For calculating the IDM, the proportion is determined biannually between the number of beneficiaries of the undergraduate education loan who are dropouts; are not active in SPADIES; and are attributed with a late payment longer than 90 days, against the total number of beneficiaries who received a loan from ICETEX. 18 Lozano Rojas F., 2018. Evaluación de Impacto del programa ACCES II Fase 2.

Año

evaluación

N°

Beneficiarios

N° Desertores de

crédito

N° Desertores de

crédito moroso

%

Deserción

crédito

%

Desertores

morosos

Índice de

deserción y

mora

2009 106.920 27.211

9.052 25,4% 33,3% 8,5%

2011 130.051 22.590

10.313 17,4% 45,7% 7,9%

2013 171.473 30.315

8.891 17,7% 29,3% 5,2%

2015 193.148 28.814

8.467 14,9% 29,4% 4,4%

2017 225.080 17.956

7.504 8,0% 41,8% 3,3%


Page 66 of 73

beneficiaries to have access to a more pertinent offer and (b) it has encouraged institutions to improve their quality standards to compete for the demand of places. In this regard, it should be noted that the enrollment in good-quality programs and institutions went from 39 percent in 2014 to 44 percent in 2017.

29. Nonetheless, more than 50 percent of the demand of loans in ICETEX accounts to nonaccredited programs. Additionally, it is evidenced that the accredited tertiary education institutions (both public and private) are mostly located in the main cities which represents an impediment for the support of beneficiaries and the institutions located in other regions and especially, in remote areas. According to MEN, by the end of 2017, 49 universities in the country out of 292 (20.24 percent) and 1,211 programs out of 11,799 (10.26 percent) were accredited with high quality. This represents another risk for ICETEX if it is considered that most of the current loans are given to private institutions.

1.4. Equity

30. ICETEX has long tradition of prioritizing access to low-income populations and, through ACCES II Phase 2, this focus increased to cover special populations. In 2017, this resulted in a participation of 24.6 percent of these populations, an increase of 10 percent with respect to 2014. Consequently, the goal for indigenous population was largely exceeded: from the total granted loans, 57 percent went to vulnerable populations, 23 percent to Afro-Colombians and 20 percent to victims of violence.

31. In 2016, as a result of the assessment of the Safeguarding Plans, a shock plan was implemented with observed results by the end of 0217. This plan included, among other actions, (a) visits to indigenous councils, (b) events with victims of violence, and (c) prioritization of mobile offices in rural areas.

32. Analysis of the surveys and the focal groups stemming from the social assessments made in 2016 and 2017 indicate that the education loans offered by ICETEX are considered a valuable and, in some cases, only alternative available for low-income populations and populations under special conditions (see Annex 2. Closing Document of the Safeguarding Plan of ACCES II Phase 2). Nevertheless, being able to increase coverage in remote areas continues to pose an enormous challenge for the entity.

Remote Areas

33. The regions with high UBN conditions have a lower number of persons in tertiary education and in the last few years they have had a high percentage of population outside the system (MEN 2015). This might be related to a low enrollment in rural area (according to MEN, 62 percent of young people from rural area do not enroll in high school), and limited offer of accredited tertiary education in these areas.

34. In accordance with DANE (2016), lack of opportunities to study made special populations to change their place of residence in the last five years. Furthermore, according to MEN (2016), in some departments, the percentage of population outside the system is higher than 87 percent, compared to the national level, which registered an average of 59.62 percent. Additionally, in accordance with the study conducted by Cárdenas (quoted in Mora 2016), the departments with more indigenous and Afro-Colombian population have a lower performance in the SABER. All of these indicate that, to increase tertiary education in remote areas, young people needs to complete high school first but at good quality.


Page 67 of 73

1.5. Institutional Strengthening

35. Financial sustainability. Assets went from US$2.8 billion in 2014 to US$4.6 billion in 2017, which is attributed to higher income for portfolio interests, recovery of written-off portfolio, recognition of subsidies of rate by the nation, and returns derived from management of investments. Additionally, the capacity to continue to leverage placement of loans is attributable to growth of profits and the efficient management of resources, as evidenced by the Return on Equity (ROE) indicators in 13.7 percent, the solvency ratio in 67 percent, and profitability of the asset in 8 percent by 2017. During ACCES II Phase 2, ICETEX had a long-term risk rating of AAA and short-term risk rating of 1+, granted by Fitch Ratings.

36. Financial innovation. The consolidation of new products has served the purpose of increasing and diversifying sources of financing. These include: the use of resources from abandoned accounts (more than COP 900 billion); the introduction of scheduled savings or ‘EduPlan’ to allow interested parties to schedule savings that will finance first semesters of tertiary education; the study of the financing contingent upon income (FCI), an initiative where beneficiaries would receive assistance in tuition and support, and when they enter to work life, they will pay according to their income; and the creation of the SER Scholarships, to collect and administer donations from the private sector.

37. Strengthening of the Education Loan Portfolio. As of December 31, 2017, a total of 404,357 obligations were registered in the active portfolio with a balance of capital of COP$4.75 trillion. With respect to 2014, the value of the portfolio grew by 53 percent. The portfolio of the lines ACCES and ACCES-Tú Eliges stands at 74 percent of the total portfolio in 2017, with a total of 290,456 active beneficiaries in the amount of $3.52 trillion.

38. Collection management. In April 2016, ICETEX assumed management of out-of-court collection of portfolios with a default higher than 90 days, which allowed establishing segmented collection strategies with a more effective and less aggressive method of collection. Moreover, ICETEX increased collection by 52.7 percent between 2014 and 2017, reaching $834,965 million as of December 2017; and managed to reduce to one digit the indicator of nonperforming loans (ICV), going from 13.19 percent in December 2014 to 8.17 percent in December 2017. If default longer than 90 days is considered, the ICV decreases from 19.29 percent in 2014 to 10.68 percent in 2017 only considering the amortization period.

39. Conciliations with tertiary education institutions. Validation of disbursements made for payment of tuition of each one of the beneficiaries has been essential to ICETEX. As of December 31, 2017, important progress was made in these matters, with minutes signed reaching 95.48 percent of success.

40. Technological modernization and transformation. The design and implementation of core banking made ICETEX into a reference of customer service in the financial market. Accordingly, in 2016 the project contracted amount totaled $19,859,215,283 and in 2017 the stages of analysis, planning, design, and parameterization were completed.

Lessons Learned and Recommendations of ACCES II Phase 2

41. The change in the approach of the tertiary education policy affected some goals. Impacts of the Program Tú Eliges on coverage of ICETEX involved an accelerated execution of the resources of the project (2016–2017). It was observed that the simultaneous focus on quality and coverage had incompatibilities


Page 68 of 73

of result, that is, with the same resources, less education loans were financed in the lines ACCES-Tú Eliges, although with a better quality and higher benefit of financing of students (100 percent tuition).

42. The need to strengthen the offer in tertiary education was evidenced. Accordingly, the role of ICETEX increases its relevance as it must promote and boost tertiary education comprehensively.

43. It is essential that continuity of the project with external loan resources contributes, on the one hand, to consolidate financial sustainability of ICETEX and, on the other hand, to integrate the necessary components to boost the offer of tertiary education in Colombia, in accordance with the guidelines of public policy. For this reason, in the design of PACES, the program Colombia Científica (Scientist Colombia) was included. This program is an important commitment with improvement of the quality of the system by boosting research, innovation, and technological development.

44. More work is required in financial education of the beneficiaries and potential beneficiaries. During implementation of ACCES II Phase 2, different actions of financial education were conducted for beneficiaries and potential beneficiaries, including implementation of modules on the web page of ICETEX and other activities, managing to have an impact on about 5,000 students in 53 universities of 14 departments. Nevertheless, in focal groups with beneficiaries, it was evidenced that loopholes in understanding the financial conditions of the education loan remain. For this reason, ICETEX must strengthen even more financial education in the short and medium term.

45. Have an appropriate team for administration of the project was key. With ACCES II Phase 2, the Project Coordination Unit was strengthened, managing to integrate a specialist for monitoring and follow-up, as well as hiring of social consultants for promotion and follow-up of the Safeguarding Plan. Likewise, the project was in charge of hiring and supervision of external audits of the financial statements of project ACCES II Phase 2, conducted by an internationally renowned independent audit firm.

46. Good management of the foreign exchange risk of previous phases was reestablished. The project continued with the process of conversion of debt as of the date of request of reimbursement. In this manner, the effective date of conversion occurs at request, minimizing the risk of exchange rate in payment of interests and financial cost of the debt. This has allowed ICETEX to have more flexibility while making the disbursement request. In this manner, it has two options of effective date of conversion, that is, based on the results of the best and market financial estimations, the optimal option is chosen.

47. Proper supervision allowed planning early warnings. Thanks to the supervision endeavors of the World Bank, including a review of the equity indicators in 2016, ICETEX implemented a plan to improve attention to vulnerable population. This allowed accomplishments in these indicators and improvement of the benefits to ethnic communities and victims of violence.


Page 69 of 73

ANNEX 6. SUPPORTING DOCUMENTS

Departamento Nacional de Planeación. 2010. Plan Nacional de Desarrollo 2010-2014. Bogota, Colombia: Departamento Nacional de Planeación.

Departamento Nacional de Planeación. 2014. Bases del Plan Nacional de Desarrollo 2014-2018. Bogota, Colombia: Departamento Nacional de Planeación.

Fack, G., and J. Grenet. 2015. Improving College Access and Success Forlow-Income Students: Evidence from a Large Need-based Grant Program. American Economic Journal: Applied Economics 7 (2): 1–34.

Ferreyra, M. M., C. Avitabile, J. Botero Álvarez, F. Haimovich Paz, and S. Urzúa. 2017. At a Crossroads: Higher Education in Latin America and the Caribbean. Washington, DC: World Bank Group.

Lozano-Rojas, F. 2018. “Effects of Cash Subsidies Embedded in Student Loans.” Working Paper.

Melguizo, T., F. Sanchez, and T. Velasco. 2016. “Credit for Low-Income Students and Access to and Academic Performance in Higher Education in Colombia: A Regression Discontinuity Approach.” World Development 80: 61–77.

Nakajima, N. 2016. “A practical guide to conducting economic and financial analyses for compulsory education projects.” Draft: 1–23.

Rau, T., E. Rojas, and S. Urzúa. 2014. “Loans for Higher Education: Does.” NBER Working Paper No. 19138.

World Bank. 2011. Country Partnership Strategy for the Republic of Colombia for the Period FY2012–FY2016. Washington, DC: World Bank Group.

World Bank. 2014. Country Partnership Strategy Progress Report for Colombia for the Period FY12–FY16. Washington, DC: World Bank Group.

World Bank. 2014. Project Appraisal Document: Colombia - Second Phase of the Access with Quality to Higher Education Project. Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/407291468240567893/Colombia-Second-Phase-of-the-Access-with-Quality-to-Higher-Education-Project.

World Bank. 2015. Colombia: Systemic Country Diagnostic. Washington, DC: World Bank Group.

World Bank. 2016. Country Partnership Framework for the Republic of Colombia for the Period FY16–FY21. Washington, DC: World Bank Group.

World Bank. 2017. Project Appraisal Document, Colombia - Access and Quality in Higher Education Project. Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/845091484227217440/Colombia-Access-and-Quality-in-Higher-Education-Project.


Page 70 of 73

ANNEX 7. TABLES AND FIGURES

Figure 1. Average Loan Disbursement over Time (COP, millions)

Source: Planning Office of ICETEX.

Figure 2. Dropout Rate by Type of Program

Source: MEN, SPADIES.

3.644.01

4.05

4.66

4.44

5.99

0

1

2

3

4

5

6

7

2015-2 2061-1 2016-2 2017-1 2017-2 2018-1

CO

P$

Mill

ion

s

22.30%

20.10%

22.79% 22.42%

19.40%18.32%

17.12%

12.90%11.80%

11.10% 10.40% 10.07%9.25% 9.03%

0%

5%

10%

15%

20%

25%

2010 2011 2012 2013 2014 2015 2016

TyT agregado Universitario


Page 71 of 73

Figure 3. Proportion of Total Enrollment by SISBEN Level (%)

SISBEN Level 2010 2011 2012 2013 2014 2015 2016

Level 1 19.69 21.82 23.49 24.93 26.34 28.22 31.47

Level 2 20.28 21.42 22.41 21.92 21.95 22.24 22.92

Level 3 5.64 5.65 6.02 5.73 5.64 5.44 5.15

Other Level 1.50 1.09 1.04 0.99 0.94 0.90 0.88

No SISBEN 52.89 50.02 47.04 46.43 45.14 43.20 39.58

Source: MEN - SPADIES.

Figure 4. Proportion of Enrollment in Higher Education by Economic Quintiles

Source: DNP, from DANE.

5.8% 6.5% 7.3% 9.1% 10.3%

10.4% 12.2% 13.5%15.0%

17.0%

21.0%21.2% 20.1%

22.9%21.6%

31.8% 28.4%30.7%

27.8% 28.0%

31.0% 31.6% 28.4% 25.2% 23.1%

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

2008 2011 2013 2015 2017

quintil 1 quintil 2 quintil 3 quintil 4 quintil 5


Page 72 of 73

Figure 5: Number of Accredited Institutions by Department/Region

Source: MEN.

Figure 6. Nonperforming Loan Ratio of ICETEX Loans

Source: Planning Office of ICETEX.

17.63%16.19%

14.86%15.61% 15.67%

13.19%12.69%

10.00%10.44% 10.90%

8.76%8.17%

0%

5%

10%

15%

20%


Page 73 of 73

Figure 7. Assets, Liabilities, and Cash Flow of ICETEX, Millions of COP$

December 2014

December 2015

December 2016

December 2017

Assets 2.856.545 3.353.324 3.817.927 4.612.568

Liabilities 962.690 1.138.704 1.345.514 1.726.425

Equity 1.893.855 2.214.620 2.472.413 2.886.143

Profit 175.999 223.161 160.116 338.542

Source: ICETEX, Financial Vice-presidency

Implementation Completion and Results Report (ICR) Document · 2020. 7. 17. · document of the...

Documents

Transcript of Implementation Completion and Results Report (ICR) Document · 2020. 7. 17. · document of the...