document of the world bank · of cities to resume execution. funding is now secure for works on the...

Document of The World Bank

Report No: ICR00003576

IMPLEMENTATION COMPLETION AND RESULTS REPORT (IBRD-75780)

ON A

LOAN

IN THE AMOUNT OF USD 35.64 MILLION

TO THE

MUNICIPALITY OF SÃO LUÍS

FOR A

SÃO LUÍS ENHANCING MUNICIPAL GOVERNANCE AND QUALITY OF LIFE PROJECT

April 26, 2016

Water Global Practice Brazil Country Management Unit 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 October 31, 2015

Currency Unit = Brazilian Real

R$ 1.00 = US$ 0.259 US$ 1.00 = R$ 3.855

FISCAL YEAR

January 1 – December 31

ABBREVIATIONS AND ACRONYMS

APL Adaptable Program Loan BMLP Brazil Municipal Lending Program BTT Bank Task Team CAEMA State Water and Sewerage Utility CBA Cost Benefit Analysis CEA Cost Effectiveness Analysis CEF Federal Loan and Savings Bank (Caixa Econômica Federal) CPF Counterpart Funds CPL Permanent Procurement Center CPS Country Partnership Strategy CSL Sector Bidding Committee EIRR Economic Internal Rate of Return ETE Sewerage Treatment Plant FMR Financing Monitoring Reports FRL Fiscal Responsibility Law IBGE Brazilian Institute of Geography and Statistics ICMS State Value Added Tax LED Local Economic Development MCMV Federal Low Income Housing Program (Minha Casa Minha Vida) M&E Monitoring and Evaluation MIS Project Management Information System MLP Municipal Lending Program O&M Operation and Maintenance PAC Federal Growth Acceleration Program PAD Project Appraisal Document PBB Bacanga Basin Program (Project) PMSL Municipal Government of São Luís SADES Sub-Secretariat for Sustainable Development SEBRAE Brazilian Service for Support to Small Business SEMA State Secretariat of Environment SEMCAS Municipal Secretariat for Child and Social Assistance SEMMAM Municipal Secretariat for Environment SEMOSP Municipal Secretariat for Works and Public Services SEMPE Municipal Secretariat for Special Projects SEMURH Municipal Secretariat for Housing and Urban Development

SEMUSC Municipal Secretariat for Security and Citizenship SEMSUR Municipal Secretariat of Urban Services SENAI National Service for Industrial Training SEPLAN Municipal Secretariat of Planning SETUR Municipal Secretariat for Tourism SIGA Environmental Management Information System SINFRA State Secretariat for Infrastructure SPGP System for Program Management and Planning SPU State Secretariat for National Assets (Patrimony) TA Technical Assistance UFMA Federal University of Maranhão UGP Project Management Unit (Unidade de Gerenciamento do Projeto) WSS Water Supply and Sanitation

Regional Vice President Jorge Familiar Calderon

Senior Global Practice Director: Jennifer Sara

Practice Manager: Wambui Gichuri

Project Team Leaders: Emanuela Monteiro/Juliana Garrido

ICR Team Leaders: Emanuela Monteiro/Juliana Garrido

COUNTRY: BRAZIL

São Luís - Enhancing Municipal Governance and Quality of Life Project

CONTENTS

Data Sheet A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Project Performance in ISRs

H. Restructuring I. Disbursement Graph

1. Project Context, Development Objectives and Design .............................................. 12. Key Factors Affecting Implementation and Outcomes ............................................. 53. Assessment of Outcomes ......................................................................................... 164. Assessment of Risk to Development Outcome ......................................................... 215. Assessment of Bank and Borrower Performance ..................................................... 226. Lessons Learned ....................................................................................................... 247. Comments on Issues Raised by Borrower/Implementing Agencies/Partners .......... 26Annex 1. Project Costs and Financing .......................................................................... 27Annex 2. Outputs by Component ................................................................................. 30Annex 3. Economic and Financial Analysis ................................................................. 42Annex 4: Safeguards Compliance and Outcomes…………………………………… 57Annex 5. Bank Lending and Implementation Support/Supervision Processes ............ 66Annex 6. Beneficiary Survey Results ........................................................................... 68Annex 7. Stakeholder Workshop Report and Results ................................................... 70Annex 8. Summary of Borrower's ICR and/or Comments on Draft ICR ..................... 71Annex 9. Comments of Cofinanciers and Other Partners/Stakeholders ....................... 81Annex 10. List of Supporting Documents .................................................................... 82

Annex 11. Photographic Record of Project Investments…………………………….. 84 MAPS

A. Basic Information

Country: Brazil Project Name:

BR Municipal APL: Sao Luis Enhancing Municipal Governance and Quality of Life Project

Project ID: P094315 L/C/TF Number(s): IBRD-75780 TF 5540

ICR Date: 11/17/2015 ICR Type: Core ICR

Lending Instrument: SIL Borrower: MUNICIPALITY OF SAO LUIS

Original Total Commitment:

USD 35.64M Disbursed Amount: USD 23.44 M (Client Connection)

Revised Amount: NA

Environmental Category: A

Implementing Agencies: Municipal Government of São Luís/Secretariat for Special Projects (SEMPE), State of Maranhão

Co-financiers and Other External Partners: N/A B. Key Dates

Process Date Process Original Date Revised / Actual

Date(s)

Concept Review: 07/28/2005 Effectiveness: 01/28/2009 01/28/2009

Appraisal: 08/02/2007 Restructuring(s): 12/12/2013

Approval: 07/17/2008 Mid-term Review: 03/19/2012 05/21/2012

Closing: 12/31/2013 10/30/2015 C. Ratings Summary C.1 Performance Rating by ICR

Outcomes: Unsatisfactory

Risk to Development Outcome: Significant

Bank Performance: Unsatisfactory

Borrower Performance: Unsatisfactory

C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings

Quality at Entry: Unsatisfactory Government: Unsatisfactory

Quality of Supervision: Moderately Unsatisfactory

Implementing Agency/Agencies:

Moderately Unsatisfactory

Overall Bank Performance:

Unsatisfactory Overall Borrower Performance:

Unsatisfactory

C.3 Quality at Entry and Implementation Performance Indicators

Implementation Performance

Indicators QAG Assessments

(if any) Rating

Potential Problem Project at any time (Yes/No):

No Quality at Entry (QEA):

None

Problem Project at any time (Yes/No):

Yes Quality of Supervision (QSA):

None

DO rating before Closing/Inactive status:

Unsatisfactory

D. Sector and Theme Codes

Original Actual

Sector Code (as % of total Bank financing)

Flood protection 52 60

General water, sanitation and flood protection sector 22 3

Other social services 4 12

Solid waste management 9 --

Sub-national government administration 13 25

Theme Code (as % of total Bank financing)

City-wide Infrastructure and Service Delivery 29 60

Improving labor markets 14 10

Other environment and natural resources management 14 8

Urban Economic Development 14 10

Urban services and housing for the poor 29 12 E. Bank Staff

Positions At ICR At Approval

Vice President: Jorge Calderon Pamela Cox

Country Director: Martin Raiser John Briscoe

Practice Manager/Manager:

Wambui Gichuri Laura Tuck

Project Team Leader: Emanuela Monteiro/Juliana Garrido

Ming Zhang

ICR Team Leader: Emanuela Monteiro/Juliana Garrido

ICR Primary Author: Anna Roumani F. Results Framework Analysis

Project Development Objectives (from Loan Agreement) The objective of the São Luís Enhancing Municipal Governance and Quality of Life Project is to improve the Borrower’s capacity in the managerial, financial, urban, environmental and service-delivery areas with a view to promoting local economic development and improving the quality of life of the population living in the Bacanga River Basin. Revised Project Development Objectives (as approved by original approving authority) N/A (a) PDO Indicator(s)

Note: The reviewed indicator is presented on the top line. In all cases where the original indicator (PAD) was revised by the Restructuring (2013), the original version is stated in the Comments section.

Indicator Baseline Value

Original Target Values (from

approval documents)

Formally Revised Target Values

Actual Value Achieved at

Completion or Target Years

Indicator 1 :

PDO Indicators: Development of the LED strategy by 2011, and budget allocated to undertakemain actions from the strategy starting in 2012.

Value quantitative or Qualitative)

Forum for Sustainable Development in Sao Luis

Adoption 2009; and, budget allocated 2010

Develop by 2011; and, budget allocated for implement. starting 2012

LED strategy developed. Budget allocated and actions undertaken.

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015

Comments (incl. % achievement)

Achieved: Activities were developed, budgeted and undertaken: (i) database with policy and decision-making socio-economic support information; (ii) capacity-building plan for municipal staff; (iii) municipal program for institutional supply and acquisition. LED Plan was basis for preparing: State’s Multi-Year Plan, 2014-2017; and Municipal Strategic Plan for Sao Luis, 2033. Section 3.2 and Annex 2. Original Indicator (PAD): Adoption of the local economic development (LED) strategy by 2009, and budget allocated to undertake main actions from the strategy starting in 2010.

Indicator 2 : Development of a competitiveness enhancement plan for one economic cluster, and implementation of recommended short-term actions from plan by 2012.


No Plan

Adopt competitiveness plan for one economic cluster; implement short-term actions by end-project.

Develop plan and implement short-term actions by 2012

Competitiveness plan for Tourism Cluster developed; short-term actions implemented.

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Achieved: 100% Competitiveness Plan for Tourism Cluster was developed and recommended short-term actions were implemented. See Main Text Section 3.2 and Annex 2. Original Indicator (PAD): Adoption of a competitiveness enhancement plan for one economic cluster, and implementation of recommended short-term actions from the plan by end of project (EOP)

Indicator 3 : Increased consumer satisfaction with water, sanitation, drainage and urban services.


No survey conducted at appraisal

Annual survey; and, aggregate 20% improvement over initial survey

End of project survey vs. previous survey

One survey conducted (baseline). Final survey still waiting procurement at ICR conclusion.

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Data not available: Satisfaction Survey not contracted due to budget constraints. However, field evidence collected from interviews with local residents and community leaders during ICR mission, as well as inspection on-site, indicate high level of satisfaction with project-financed works/services: restructured Coroado and Rio das Bicas Canals are effectively capturing rainfall and preventing flooding; paved areas, plazas and urban parks in surrounding areas with proper lighting and communal/leisure facilities are stimulating housing and business renovation. See photos Annex 11 . Original Indicator (PAD): Annual survey of population satisfaction with quality of public services conducted and aggregate population satisfaction level improved by 20% compared to the initial survey. Numerical target was dropped and the public services involved were specified. The issue in 2013 was lack of consumer surveys for 2009-2011, and baseline not conducted until 2012.

Indicator 4 : Project targeted areas acquire macro drainage capacity to handle an intense storm every 25 years and micro drainage capacity for a storm every 5 years


No capacity

Macro-drainage for 25-year storm; micro-drainage for 5-year storm

See comments

Capacity improved: macro-drainage works in Das Bicas and Coroado areas concluded

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015 Comments (incl. % achievement)

Substantially achieved: Macro-drainage works – the Rio das Bicas and Coroado Canals – were completed by closing. Both canals were rehabilitated structurally to a standard able to handle a storm of an intensity probable on average every 25

years (based on meteorological records), cleared of all obstructions. Micro-drainage works were incomplete/stalled but represented a small portion of the overall drainage works. See Main Text Section 2.2. Indicator targets are the same as the PAD but Restructuring in 2013 modified the phasing of achievements to reflect extension of the closing date to 10/30/2015.

Indicator 5 : Increase in percent of sewage collected (from 39% to 80%) and treated (from 0% to 80%) in the targeted area.


39% collected and 4% treated

Sewage collected 80%; sewage treated 80%

Sewage collected 80%; sewage treated 41%

No increase. Sewage collected 39% and treated 4% as per baseline

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Not achieved: Sewerage works, to be fully-financed by counterpart funds (PAC), were contracted and initiated but halted due to design inconsistencies. Borrower contracted a consultant to revise the designs which are now ready for procurement.At closing, SEMPE and CAEMA were seeking funding from the Federal Ministry of Cities to resume execution. Funding is now secure for works on the Left Bank. CAEMA’s sewerage works on Right Bank are 30% complete and still in the budget pipeline. Targets were modified by Restructuring (2013) to reflect the Project’s actual investments in sanitation for completion by 2015, and implementation status at the time of Restructuring.

Indicator 6 : 100% of families threatened with flooding from regular tidal movements of the Bacanga Lake in Project targeted area are relocated to the satisfaction of residents.


Zero 100% of families 100% of families

18% of families

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Partially achieved: 18% 108 families were voluntarily resettled in housing units (apartments) in the Piancó VII and VIII clusters – financed by the Federal Minha Casa Minha Vida low-income housing program – from an updated estimate of 568. Another 417 cadasters were under review by the Bank of Brazil, the responsible financial institution, for settlement in Piancó I, II and III housing developments, of which 260 were approved and are expected to be resettled by end-May 2016. No beneficiary/resident results are available due to lack of the planned Satisfaction Survey. The Restructuring modified the phasing of achievement to reflect extension of closing date to 10/30/2015.

Indicator 7 : Direct project beneficiaries (#) of which female (%)


Zero 92,500 WSS and drainage; 50% female

19,513 direct project beneficiaries from macro-drainage works; 9,952 women (51%)

Date achieved 12/10/2013 10/30/2015 10/30/2015


Exceeded for macro-drainage: 130% (female 51%); zero for WSS: Results measured include Coroado Canal (7,241 persons) and Rio das Bicas/Salinas do Sacavem Canal (12,272 persons). Water supply and sanitation investments were not completed. There were also 141 families resettled before closing, and 5,420 families benefited by parks and plazas (mostly in the vicinity of the two canals). This was a new Core Indicator added by Restructuring (2013), and intended to measure only WSS and drainage beneficiaries. Borrower revised the number of beneficiaries from infrastructure improvements to 92,500 (77,500 from WSS,15,000 from macro-drainage works).

(b) Intermediate Outcome Indicator(s)

Indicator Baseline Value

Original Target Values (from

approval documents)

Formally Revised

Target Values

Actual Value Achieved at

Completion or Target Years

Indicator 1 : Component 1: Elaboration of Local Economic Development strategy

Value (quantitative or Qualitative)

Permanent Forum for Sustainable Development of São Luís

Main actions from the strategy adopted

DROPPED NA


Dropped by Restructuring because PDO Indicator #1 was similar.

Indicator 2 : Establishment of mechanisms for critical clusters to promote collaboration and joint actions


Workshops with stakeholders

Cluster organization a critical factor in member collaboration

DROPPED NA

Date achieved 10/01/2008 12/31/2013 12/10/2013 10/30/2015


Dropped by Restructuring. Revised PDO Indicator #2 already had the means to monitor the progress of the work contracted (Cluster Development Plan) and its result. Also, PDO Indicator #3 provided milestones to accompany the intermediate results of activities related to support for critical clusters.

Indicator 3 : Reduced number of procedures to start a business Value (quantitative or Qualitative)

18 12 DROPPED NA


Dropped by Restructuring in alignment with the proposed scope changes.

Indicator 4 : Number of micro-enterprises assisted in Project area


5 70 540 540

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Achieved: 100% Training was provided to micro-entrepreneurs under the “Plano de Inovação” in entrepreneurship, pricing, sales, HR management and quality control. See Section 3.2 and Annex 2. Targets were modified by the Restructuring to reflect actual project investments and implementation status at that time.

Indicator 5 : Number of people trained in the Project area Value (quantitative or Qualitative)

100 2,000 people (cum. by 2012)

676 people 676 people

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Achieved: 100% Training was delivered to 676 people by SENAI and SEBRAE in basic job skills (including civil construction, masonry, iron-working, electrical and painting) under the “Plano de Inovação”. See Section 3.2 and Annex 2 Targets were modified by Restructuring to reflect actual project investments and implementation status at that time.

Indicator 6 : Formulation and execution of annual work plan by the Secretaria Adjunta de Geração de Emprego, Trabalho e Renda (SAGETR)


Annual Work Plan prepared

Annual Work Plan prepared

DROPPED NA

Date achieved 07/17/2008 12/31/2013 12/10/2013 10/30/2015


Indicator was dropped by Restructuring: SEMPE could not be held accountable for ensuring the execution of annual plans of another Secretariat; all Municipal Secretariats are responsible for developing and implementing own planning instruments; and, activities of SAGETR were absorbed by the Sustainable Development Sub-Secretariat of SEPLAN, a Project partner.

Indicator 7 : Development and implementation of actions for strengthening the Borrower’s procurement system.


30% Under 5% Actions developed and implemented

Actions developed and implemented

Date achieved 07/17/2008 12/31/2013 12/10/2013 10/30/2015


Achieved: 100% Actions were developed and implemented, including training of procurement staff within SEPLAN and in the PMSL’s Permanent Procurement Center. Original Indicator (PAD): Reduction of the percentage of procurement responses that exceed the 45-day limit. The Borrower requested that this be dropped in favor of an indicator which could monitor project investments in strengthening the Municipal procurement program.

Indicator 8 : Component 2:

Percent of sewage collected Value (quantitative or Qualitative)

39% 80% 80% 39% (no change from baseline)


Not achieved: This indicator replicates part of PDO Indicator #5. See details regarding sewage collected in Comments section of that indicator.

Indicator 9 : Percent of total sewage treated Value (quantitative or Qualitative)

4% 80% 41% 4% (No change)

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Not achieved: Sewerage works were stalled at the time of ICR preparation. See PDO Indicator #5 for details concerning sewage treated. As per PDO Indicator #5, targets were modified by the Restructuring to reflect Project’s actual investments in sanitation at that time, the revised closing date and then-current implementation status.

Indicator 10 : Percent of intervention area flooded in 5-year rain Value (quantitative or Qualitative)

40% (Correct Baseline is 10%. See Comments)

0% 0% Not achieved

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Not achieved: This indicator refers to an intensity probability of 5 years and merited clarification. Micro-drainage works were not completed as inter alia, they were linked to counterpart-funded street paving works. However, micro-drainage works were a small portion of the overall drainage intervention and, the rehabilitated macro-drainage system has, so far, captured heavy rainfall and prevented flooding. Indicator was revised by Restructuring. The baseline/target was adjusted due to having been switched in error at appraisal with #11 below; and, the phasing of achievements was revised due to extension of the closing date.

Indicator 11 : Percentage of intervention area flooded in a 25-year rain Value (quantitative or Qualitative)

10% (Correct Baseline is 40%. See Comments)

0% 0% Not available

Date achieved 01/2008 12/31/2013 10/30/2015 10/30/2015


Data not available: This indicator refers to an intensity probability of 25 yearsand merited clarification. UGP lacked capacity to adequately monitor this indicator which required a specific M&E structure (i.e., understanding of the assumptions underlying the detailed engineering designs; monitoring/follow-up of meteorological data on rains; and, capacity to measure impact in the field). As a proxy, the Rio das Bicas and Coroado Canal macro-drainage works intended to control flooding in the intervention area were completed and, local residents interviewed by the ICR missions attested that heavy rains since completion, which would normally provoke flooding, had not done so.

The Restructuring revised the baseline/target due to its having been switched in error at appraisal with #10 above; and, the phasing of achievements was revised due to extension of closing date.

Indicator 12 : People in urban areas provided with access to “Improved Water Sources” under the Project


Not available/zero 77,551 Zero

Date achieved 12/10/2013 10/30/2015 10/30/2015


Not achieved: At project closing, about 350 connections had been installed using counterpart (PAC) funding. However, the water production system (CAEMA’s responsibility) still lacked the capacity to feed the network and thus the connections were non-functional. Works were stalled at closing and the schedule for completion is not clear. The PMSL stated that their completion is a top priority, final engineering designs are ready but dependent on the release of approved Federal funds to proceed with contracting. This was a new (Core) Indicator added by Restructuring. Targets were established based on proxy: # connections to be provided multiplied by average # of people per family (3.6).

Indicator 13 : People in urban areas provided with access to “Improved Sanitation” under the Project


Not available/zero 73,840 people Zero

Date achieved 12/10/2013 10/30/2015 10/30/2015


Not achieved: See PDO Indicator #5 for background. This was a new (Core) Indicator added by Restructuring. Targets were established based on proxy: # connections provided multiplied by average # people/family (3.6).

Indicator 14 : Number of families living in flood-prone areas benefiting from improved drainage


Not available/zero 3,800 families 4,166 families5,420 families (area of Coroado and Das Bicas Canals)


Exceeded: 130% This result measures beneficiaries of improved drainage in the area surrounding the Coroado and Das Bicas Canals. Detailed methodology for estimating beneficiaries was provided by UGP.

Indicator 15 : Percentage of area households receiving 24-hour water supply services Value (quantitative or Qualitative)

10% 100% DROPPED NA

Date achieved 07/17/2008 12/31/2013 12/10/2013 10/30/2015


Restructuring determined that this indicator was inappropriate. Project water supply (WS) investments would be unable to reach such targets. Only after investments in the Italuís II Water Supply System (investments not linked to the Project), would CAEMA be able to supply WS services 24/7.

Indicator 16 :

Component 3: Percentage of families living in flood-prone areas become free of flood threat


0% 100% DROPPED NA


See PDO Indicator #6 referring to the impact of resettlement.

Indicator 17 : Satisfaction of resettled population


NA Survey conducted and results satisfactory

Survey conducted and results satisfactory

Final Satisfaction Survey unavailable

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Unknown: Due to budget constraints, the Satisfaction Survey was still awaitingprocurement at the time of ICR conclusion. See Annex 11 for photos of housing quality before and after resettlement. The ICR mission met with residents, and inspected the buildings/grounds, and interior of an apartment. Prima facie evidence suggests strong satisfaction with new living conditions. Restructuring revised the indicator to reflect then-current implementation status and the extended project closing date.

Indicator 18 : Non-rural roads rehabilitated Value (quantitative or Qualitative)

Zero 39 km paved 62 km paved 4.7 km paved

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Negligible: 7.6% The 4.7 km completed by closing represent the paving of 33 stretches of road. Remaining works were stalled, awaiting revised engineering designs. Additional funding was being sought and a new procurement process will be launched. Original Indicator (PAD): Paving of 39 km of road The Indicator was revised to a Core equivalent designed to reflect the Project’s actual investments fully counterpart-financed under the Pro-Transportes Federal program, as well as implementation status at that time.

Indicator 19 : Bacanga Dam floodgates completely repaired and operated effectively


Floodgates operating inadequately

Floodgates operating properly

Bacanga Dam repaired and floodgates operating properly

Floodgates not repaired or operated effectively.


Not achieved: Floodgates continued to be operated improperly, rehabilitation works did not occur and the main floodgate of the Dam collapsed in September

2015. The State Government, responsible for the Dam, conducted emergency works. The status of future rehabilitation is uncertain. Section 2.2 and Annexes 2and 4. Indicator was revised by Restructuring. Target was modified to reflect previous delays and the then-current implementation status/schedule for dam rehabilitation.

Indicator 20 : Operations manual developed for the floodgates Value (quantitative or Qualitative)

Zero Operational Manual for floodgates

DROPPED NA

Date achieved 07/17/2008 12/31/2013 12/10/2013 10/30/2015


This indicator was poorly-designed at appraisal and not considered relevant. Indicator #19 is considered as taking into account dam operation, technical assistance, and guidelines/rules for O&M during and after dam rehabilitation works.

Indicator 21 : Integrated water, drainage, sanitation and solid waste plan developed and adopted.

Value (quantitative For Qualitative)

No plan Plan developed Plan developed and adopted

Integrated plan was developed and public consultations were held. Plan not yet submitted to the municipal legislature for approval.

Date achieved 07/17/2008 12/31/2013 10/30/2015 10/30/2015


Partially achieved: The Plan was developed and publicly-consulted but, at ICR conclusion, remained with the Municipal Attorney General, responsible for conveying the draft law to the municipal legislature for approval. Original Indicator (PAD): Solid waste treatment plan developed. Indicator was revised in line with scope changes under the Restructuring (2013).

Indicator 22 : Target population with use or ownership rights recorded as a result of the Project


Zero 3,000 people Zero

Date achieved 12/10/2013 10/30/2015 10/30/2015


Not achieved: The Borrower informed the Bank that the Federal University of Maranhão (UFMA), in coordination with the State Government, is working on land regularization in the area targeted. See matrix, Annex 2. This was a Core Indicator added by Restructuring (2013).

Indicator 23 : Public space (m2 of green areas, urban parks) created as a result of the Project


Zero 83,585.50 m2 Total of 17,855 m2 of green areas and urban parks created

Date achieved 12/10/2013 10/30/2015 10/30/2015


Partially achieved: 21.4% This comprised 6,372 m2 (5 Praças under the Right Bank urbanization – Rio das Bicas Park); 8,553 m2 (3 Praças under Left Bank urbanization – requalification of Left Bank/Lot 1); and, 2,930 m2 of public, paved road and space surrounding the Rio das Bicas Canal. The Restructuring foresaw a more comprehensive intervention. See Annex 2 matrix.

G. Ratings of Project Performance in ISRs

No. Date ISR Archived

DO IP Actual

Disbursements (USD millions)

1 10/09/2008 Satisfactory Satisfactory 0.00 2 04/23/2009 Satisfactory Satisfactory 2.06 3 11/19/2009 Satisfactory Satisfactory 2.06 4 04/06/2010 Satisfactory Satisfactory 2.06 5 12/22/2010 Satisfactory Satisfactory 2.77 6 06/26/2011 Satisfactory Moderately Satisfactory 3.97 7 02/01/2012 Satisfactory Moderately Satisfactory 6.67

8 11/18/2012 Moderately SatisfactoryModerately

Unsatisfactory 13.25

9 07/22/2013 Moderately

Unsatisfactory Moderately


10 03/07/2014 Moderately SatisfactoryModerately


11 10/19/2014 Moderately

Unsatisfactory Unsatisfactory 17.70

12 04/01/2015 Unsatisfactory Moderately






H. Restructuring (if any)

Restructuring Date(s)

Board Approved

PDO Change

ISR Ratings at Restructuring

Amount Disbursed at

Restructuring in USD millions

Reason for Restructuring & Key Changes Made

DO IP

09/02/2010 -- S S 2.50 To reflect Municipal Government’s access to additional counterpart funding

1 Client Connection shows US$23.44 m disbursed at the time of ICR finalization. The above is disbursement as at project closing (10/30/2015).

Restructuring Date(s)

Board Approved

PDO Change

ISR Ratings at Restructuring

Amount Disbursed at

Restructuring in USD millions

Reason for Restructuring & Key Changes Made

DO IP

for specific activities; and, to accelerate disbursement of the Loan. Change: Bank’s percentage share of project expenditures increased for Categories 1, 2 and 3 from 55% to 100%.

12/10/2013 -- MS MU 15.67

To reflect agreed changes, accelerate project execution andincrease likelihood of achieving PDO. Change: revised Results Framework; extended closing date 22 months; reduced project scope; changed financing plan; changed institutional arrangements; and, reallocated funds.

I. Disbursement Profile

1

1. Project Context, Development Objectives and Design

Introduction: The São Luís Enhancing Municipal Governance and Quality of Life Project (known locally as Programa Bacia Bacanga) was another in the series of Brazil Municipal Lending Program (MLP) operations. It was designed to improve - through integrated, multi-sector urban development actions - the living conditions of low-income residents of the Bacanga Basin area, increase the Municipal Government’s capacity for service delivery, fiscal management and environmental stewardship, and improve the conditions for local economic development. The Project was prepared between 2005 and 2007, approved by the Board in 2008 and became effective in early 2009. It was restructured in 2013 and closed in October 2015. Despite a 22-month extension of the closing date, key Project activities remained incomplete/stalled at closing, and one-third of the Loan was cancelled.

The Project was complex, risky and not well-aligned to institutional capacity. Its level of difficulty was under-estimated at appraisal, the engagement period was too short and the Project was not ready to implement despite its protracted and comparatively costly preparation. It is important to understand the current context for urban and water projects. Brazil is 85% urbanized and poverty is concentrated in urban areas. Massive migration from rural to urban areas has come at the cost of poor planning, lack of basic services and environmental vulnerability. Urban areas suffer from multiple and multi-dimensional problems and today house 60 percent of the nation’s poor, up from 50 percent at appraisal (see below). The urban “space” is therefore a priority if the Bank is to focus on its “twin goals”. The complexity and risks of such projects are unlikely to diminish, and the Bank will need to continue to explore new and more effective ways to support urban poverty alleviation and shared prosperity. In a municipality like São Luís, this Project shows that the Bank should not withdraw from financing such operations but needs to re-think its strategy, providing a 10-15 year, phased engagement with a baseline and progressive interventions acceptable to a succession of administrations. It should also be understood that the transformational impact of such operations is not short- or even medium-term. More complex than its Brazil MLP peers, this Project offers important lessons for improving the approach to urban development interventions.

1.1 Context at Appraisal

1.1.1 When the Project was appraised in 2007 over 80 percent of the Brazilian population lived in urban areas, generating 90 percent of national Gross Domestic Product (GDP) and accounting for 50 percent of the country’s poor. Poverty was increasingly urban but municipalities lacked the fiscal space to expand coverage of basic services while compulsory social expenditures were absorbing an increasing portion of municipal budgets. The over-riding challenge was to make Brazilian cities more livable, equitable and competitive. Investment and programs targeted to the urban poor sought to reduce functional inequality.

1.1.2 São Luís Municipality has a population of 900,000 people and is the capital of the State of Maranhão, one of Brazil’s poorest and least-developed states, ranked last or close to last in per capital household income and in access to services and health indicators. The city of São Luís itself has a significant industrial sector, a modern, strategically located seaport, Unesco World Heritage Site status and substantial potential for eco-tourism. At appraisal however, 50 percent of all heads of household in São Luís earned between one and two minimum salaries (one minimum salary was about US$200/month at that time) and 10 percent were destitute. Despite unemployment of around 22 percent, the city continued to attract migrants – many from rural areas and searching for a better life - to crowded peripheral settlements in low-lying areas along estuarine waters. These are low-

2

income and risk-prone settlements with inadequate housing, water supply, sanitation, drainage and solid waste management services. 1.1.3 Bacanga River Basin: The focus of Project activities, the Bacanga Basin is a large catchment area of 132 km2, important to economic activity, and housing a large percentage of the municipality’s low-income, peri-urban population in settlements which successive municipal governments have targeted for urban upgrading. To reduce large tidal variations in the Basin and permit the consolidation of upstream urban infrastructure, a dam and floodgate were constructed in the 1970s, creating a large saline lake and exposing previously submerged land which was subsequently occupied by informal housing prone to extreme drainage and wastewater problems due to the high water table. The dam’s floodgates were in disrepair and poorly-operated, leading to flooding in the informal border settlements, degraded water quality from uncontrolled domestic and industrial pollution and a range of other environmental hazards. 1.1.4 Rationale for Bank assistance: The Project was one of eight similar Bank-supported municipal operations under the Brazil Municipal Lending Program (BMLP) APL approved by the Board in April 2007 and designed to demonstrate how policies effectively implemented at the local level could improve the lives of the urban poor, and promote sound local governance and economic competitiveness. The Project reflected three of the four pillars of the Bank’s Municipal and City Strategy for Brazil: improving municipal, fiscal and administrative management; increasing competitiveness through local economic development; and, strengthening municipal service delivery. Support to São Luís was strategically important for the Bank given it was the capital of Brazil’s poorest state. The Project was consistent with the Government’s Multi-Year Plan, and supported the PMSL’s strategy to improve public services, quality of life of the poorest and the Municipality’s economic competitiveness. It was also aligned to the Bank’s 2004-2007 Country Assistance Strategy (CAS) – recognizing the importance of the sub-national level in economic growth and social equity – and the CAS Progress Report (2006).

1.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) 1.2.1 The original PDO (using the more complete Loan Agreement version) was “To improve the Borrower’s capacity in the managerial, financial, urban, environmental and service-delivery areas with a view to promoting local economic development and improving the quality of life of the population living in the Bacanga River Basin”. Key Indicators: (i) adoption of the local economic development strategy by 2009 and budget allocated to undertake main actions by 2010; (ii) adoption of a competitiveness enhancement plan for one economic cluster, and implementation of recommended short-term actions from the plan by end of project; (iii) annual survey of population satisfaction with quality of public services conducted and aggregate population satisfaction level improved by 20 percent compared to the initial survey; (iv) project targeted areas acquire macro-drainage capacity to handle an intense storm every 25 years and micro-drainage capacity for a storm every five years; (v) increases in the percentage of sewage collected (from 39 percent to 80 percent) and treated (from 0 percent to 80 percent) in the targeted area; and, (vi) 100 percent of families threatened with flooding from regular tidal movements of the Bacanga Lake in project-targeted area are relocated to their satisfaction.

1.3 Revised PDO (as approved by original approving authority) and Key Indicators, and reasons/justification 1.3.1 The PDO was not revised. The Bank task team used the simpler Project Appraisal Document (PAD) version and saw the need to adjust only the indicators. The Level II Restructuring

3

in late 2013 revised the PDO Key Indicators as follows: (i) Added a new Core Indicator “Direct project beneficiaries (#) of which female (%)”; (ii) “Adoption” was substituted by “development” for the 1st indicator, and the target years advanced to 2011 and 2012 (“development of the local economic strategy by 2011 and budget allocated to undertake main actions from the strategy by 2012”); (iii) “Adoption” was substituted by “development” for the 2nd and target year moved to 2012 (“development of a competitiveness enhancement plan for one economic cluster and implementation of recommended short-term actions from the plan by 2012”); (iv) Indicator #3 revised to “increased consumer satisfaction with water, sanitation, drainage and urban services”, more logical and measurable with targets modified; (v) Indicator #4 target was reduced to reflect then-current implementation status; (vi) Target for “sewage treated” under Indicator #5 was reduced from 80 percent to 40 percent as more realistic/achievable; and, (vii) Indicator #6 target was reduced from 750 families to 595 consistent with the approved Resettlement Action Plan and 2012 updated beneficiary survey. See Data Sheet and Section 2.3.

1.4 Main Beneficiaries

1.4.1 The primary project target group in the Bacanga Basin included eight neighborhoods (bairros - Via Embratel, Campo UFMA, Sá Viana, Pindorama, João Paulo, Filipinho, Coroadinho Sacavém with Praia Grande added later) housing 230,000 people with the following characteristics: 90 percent earned <two minimum salaries/month (about US$400); poor rainwater drainage aggravated by improper waste disposal causing high pollution and public health hazard; <40 percent of the population covered by sewerage collection and all sewage untreated; per capita water consumption <40 percent of the municipal average; and, inadequate sanitation contributing to declining water quality of the Lake inside the Bacanga Dam. The PAD Results Framework (RF) quantified 3,800 families living in flood-prone areas and expected to benefit from investments in urban drainage. The 2013 Restructuring added a Core Indicator targeting 92,500 direct project beneficiaries (77,500 water supply and sanitation (WSS), and 15,000 from drainage works) of which 50 percent or 46,250 to be women. The Project would also resettle 750 families in flood-risk/related areas, train 676 people in the Project area, and assist 540 micro-enterprises.

1.5 Original Components (as approved) 1.5.1 The project constituted the first phase of planned interventions in the Bacanga Basin and was to serve as a pilot for a broad-based integrated approach to urban management in under-developed areas of the Municipality. Components/activities were as follows: Component 1: Local Economic Development and Municipal Management Strengthening (est. total cost US$ 6.27 m of which Loan resources 60%) with four sub-components:

Local Economic Development (LED) and Competitiveness Strategy: LED strategy financed by the Cities Alliance; competitive upgrading strategy for Critical Clusters (Local Productive Arrangements); and, streamlining of business processes (Doing Business).

Capacity building for job and income generation support: training; facilitating access to existing micro-finance programs; technical assistance (TA) to the Secretariat for Employment and Income Generation.

Financial management enhancement: strengthening the municipal procurement system including assistance to the Permanent Procurement Center’s Central Bidding Committee.

Project implementation support: strengthening the Project Management Unit (UGP); supporting project management, monitoring and evaluation; public service satisfaction surveys; and, outreach, communication and dissemination.

4

Component 2: Sanitation and Water Improvements (est. total cost US$34.3 m of which Loan resources 59.5 percent) with four sub-components:

Sewerage System improvements: new sewage collection system for the Left Bank, feeding into the existing Bacanga Sewerage Treatment Plant (ETE); new wastewater collection system in areas not covered by CAEMA; mechanical sludge drying system at ETE Bacanga.

Storm Drainage System improvements: micro- and macro-drainage for the Coroado and Rio das Bicas Canals, including construction of storm drain networks on the Right Bank; improve hydraulic operation of urban drainage canals by cleaning/clearing, and improved operation and maintenance (O&M); rehabilitation of environmentally valuable areas to improve quality of water flowing into Bacanga Lake and to reduce decay on Right Bank.

Water Supply System improvements: expand water distribution system toward non-serviced areas, especially Left Bank; increase metered connections; equipment to control unaccounted-for water; rehabilitation of concrete structure and metal bridge supporting 900 mm distribution pipes on the Bacanga Dam; and, a Loss Control Pilot Program (CAEMA).

Water Supply and Sanitation (WSS) management strengthening and construction supervision: TA to strengthen sanitation management and supervision of construction works, including helping PMSL regulate water supply and sanitation (WSS), improving O&M of drainage services, and developing a Solid Waste Management Plan and related pilot in the Bacanga Basin.

Component 3: Urban and Environmental Improvements (est. total cost US$15.77 m of which Loan resources 60 percent) with three sub-components:

Informal area upgrading and resettlement: studies to establish legal limits on urban occupancy in the Bacanga Basin and to address accommodation of future migrants to São Luís; resettlement of families living in risk-prone areas; construction of local roads to facilitate public services access/supply; construction of public spaces and community facilities in project resettlement locales; TA for land tenure regularization; other urban improvements to enhance access and living conditions.

Rehabilitation of the Bacanga Dam: rehabilitation of the structure and electro-mechanical equipment; operational rules for the Dam balancing flood control and environmental objectives; monitoring lake water quality to support pollution control activities; investigation of Cadmium discharge; and Plan to Control Emissions.

Municipal environmental management support: implementation of the Municipal Environmental Policy; structuring an environmental licensing system; formulation and implementation of an Environmental Education Plan for Bacanga Basin in areas of high environmental degradation and poor living conditions; implementation of management plan for Bacanga State Park to control illegal settlement and land degradation.

1.6 Revised Components 1.6.1 Project components were revised by a Level II Restructuring in December 2013:

Component 1: Simplification of the Borrower’s business processes was dropped in favor of activities stemming from the LED strategy: database with policy and decision-making socio-economic information; capacity building plan for municipal staff; and, municipal program for institutional supply and acquisition.

Component 2: (i) Investments in a sludge-drying system in the ETE were dropped. The Borrower secured counterpart (CP) funds to design its expansion for execution by CAEMA; (ii) Pilot for solid waste management was dropped/transferred to the Municipal Secretariat for Infrastructure and Public Services (SEMOSP), already implementing eco-

5

points for solid waste collection in the Right and Left Banks; (iii) Rehabilitation of water distribution infrastructure connected to the Batatã Dam lost priority and was dropped; and, (iv) Borrower’s responsibility to regulate, operate and maintain the WSS and drainage services was dropped to focus on implementing project civil works and developing an Integrated Municipal Plan for WSS, drainage and solid waste.

Component 3: The environmental plan for the Bacanga State Park was dropped as the State Government had already developed such Plan using its own resources.

1.7 Other significant changes 1.7.1 There were two formal restructurings: (i) Amendment of the Loan Agreement dated September 2, 2010, increased the Bank’s share of project expenditures for Categories 1-4 from 55 percent to 100 percent to accelerate disbursement (Aide Memoire, September 14, 2010 and 2.2.14); and, (ii) Level II Restructuring (1.6.1 and 2.2.10) dated December 10, 2013: changed components/sub-components; revised the Results Framework (RF); dropped the requirement that UGP contract an Implementation Support Consultant to boost project management; revised project costs (expanding the PMSL counterpart contribution from US$23.76 m to US$50.73 m, an increase of 113 percent); reallocated US$8.29 m among categories; and, extended the closing date by 22 months to October 31, 2015. 1.7.2 Several appraisal estimates were revised: (i) the original target of 70,000 beneficiaries for investments in WSS and drainage was increased to 92,500; (ii) families expected to be resettled as a result of flood and other risks on the Left Bank were reduced to 595 by a new 2012 survey and updated Resettlement Action Plan, down from 750, and then to 568 (289 in flood risk and 279 in other areas). 1.7.3 In the period between appraisal and negotiations, without Bank knowledge, the Borrower unofficially changed the Project’s technical, conceptual and environmental approach to urbanization based on studies designed for this purpose, and then sought to implement same after effectiveness. The revised approach was conceptually inconsistent with the project idea of clearing low-lying, flood-prone, occupied areas of the Left Bank, and changed the Project’s scope (see 2.2.5 and footnote 7).

2. Key Factors Affecting Implementation and Outcomes

2.1 Project Preparation, Design and Quality at Entry 2.1.1 Preparation was supported by a PHRD Grant (TF5540, US$800,000, fully-disbursed) which financed consultancies, studies and diagnoses to determine optimal alternatives for interventions, technical approach and benefits. Total preparation cost through appraisal was US$1.21 m (Trust Funds and Bank budget). The Borrower spent significant additional, own funds. The effort was protracted due to the client’s inexperience. The Project used a holistic approach to address the main urban and environmental questions in the Bacanga Basin and its sustainable occupation. Urban works would be complemented by community-oriented investments – parks, plazas, sports and walking areas – to boost sustainability and in some locations to prevent the re-invasion and/or informal occupation of open areas. Category A Safeguards status entailed preparation of specific assessments, analyses and plans. 2.1.2 Two instruments would ensure that the State Government repaired/rehabilitated the Dam and that CAEMA executed the water and sanitation systems: (i) an agreement of June 10, 2007

6

between the State Government and PMSL legally committed the former to repair the Dam floodgates and supporting structures and to operate and maintain the Dam based on a new manual of procedures. A Dam Safety Panel of Experts established during Project preparation had identified key issues affecting the Dam’s operation, safety and environmental conditions; and (ii) an amendment to the existing Concession Contract for the provision of WSS services signed between PMSL and CAEMA on April 7, 2008, committed CAEMA to: rehabilitate the Sacavém water supply system, ensuring exclusive service provision in selected areas of the Bacanga Project area; and, to receive wastewater from the Left Bank and treat it at the existing treatment plant, to be expanded by CAEMA using Federal funds. Specific issues influencing quality at entry: 2.1.3 The Mayor and PMSL promoted the Project and its integrated vision widely among municipal secretariats, generating strong interest in participating given the unique opportunity represented by a World Bank loan. Scope expanded as urban development secretariats, public service providers and the newly-established SEPLAN competed for access to project “seed” money for a longer-term, larger set of municipal needs. The Bank sought to foster collaboration by accommodating the PMSL/secretariats, resulting in a complex, multi-sector operation.2

2.1.4 The Project’s physical design made sense per se and reflected a defined set of priorities, but was too demanding in practice for a single, five-year operation and existing capacity. The quality of the activities themselves varied, with the hard works supporting improved quality of life the most appropriate and those supporting improved public sector management, the least. Some preparation analyses lacked depth. Multiple, linked activities/institutions required exceptional technical and coordination capacity/arrangements, unwavering commitment and problem-free execution to finish on time. See 2.2.2. 2.1.5 Despite preparation exceeding three years and the expenditure of US$1.21 m, the Project was not ready to implement and its real costs were unknown. While there was complete clarity as to each aspect of design – the Project was not a “framework” operation – detailed engineering designs were left to the post-effectiveness period. Infrastructure interventions were planned and costed at the conceptual, basic level using thematic diagnoses taking into account mainly hydro-dynamic and water quality studies of the Bacanga Lake and capacity to control flooding in marginal areas. Loan financing allocated at appraisal was sufficient only for engineering designs, consultancies and works supervision, not the works per se, in key interventions such as water supply. 2.1.6 Project quality was compromised by under-estimation of the potentially negative effects of party-politics – especially strong in the State of Maranhão – the phasing of State and Municipal elections and likely non-alignment of the State and Municipal governing parties. Implications included: willingness of a new State Government to honor the Bacanga Dam Agreement which it had not signed; stability of the UGP’s leadership and technical team which in Brazil invariably rotate following elections; and, potentially, the Project’s relationship with/influence on CAEMA, a State agency responsible for independently-financed, core Project works and, on the State Secretariat of Environment (SEMA), responsible for the Bacanga State Park Management Plan.

2 The Project Concept Note questioned whether scope should be limited by not including significant governance and social components/activities, to streamline preparation and execution; and, how the trade-off of the proposed multi-sector/multi-secretariat approach – with its cost and slow-down effects – could be optimized. Concept Review Meeting Minutes (August 1, 2005) stressed the centrality of the governance/public sector management component in supporting the PMSL’s development agenda and longer-term sustainability of project interventions, not commenting on complexity.

7

2.1.7 The project’s integrated design locked inter-dependent activities, institutions, financing and contracts into a framework where unforeseen difficulties or the delay/failure of one element could compromise others and in practice, this did happen. This was also believed to make the Project difficult to restructure, limiting the options for change unless stakeholders (including the Bank) were willing to accept radical re-design. Further, neither the PAD nor the Manual of Operations (MOP) discussed the sequencing of activities; it seems the Project was to advance on all fronts simultaneously.3 2.1.8 Dependence of the PDO’s quality of life objectives on activities which were wholly counterpart-financed and needed to source funding from Federal Government/other programs introduced risks associated with a separate “category” of complementary investments with differentiated financing and approval processes, timetables and levels of definition. The Project would be dependent on those activities but lack control/influence if those resources were restricted for any reason or the institutions involved did not perform. 2.1.9 Lessons of earlier operations: Lessons cited in the PAD were quite limited/selective and were reflected unevenly in Project design. Lessons of similar MLP projects were not yet available. For example: (i) experience showed that strengthening municipal capacity worked best when linked to investment activities that leveraged the implementation of reforms (the premise underlying the structure of the PDO). Project design heeded this lesson only partially. Direct efforts to strengthen municipal capacity (Component 1) were small “seed” investments not linked directly to the investment activities. Other capacity building integrated in Components 2 and 3 (training, technical assistance, systems development) depended on the completion of their associated interventions; and (ii) the lessons advised against a “Christmas tree approach” burdening involved agencies, focusing instead on the most critical municipal capacity issues (local economic development and revenue/expenditure management) and investing in public services in a defined area. In practice, the Project included multiple, inter-dependent interventions in the defined area and, based on the PDO, managerial capacity-building goals went far beyond LED and revenue/expenditure. 2.1.10 Risks and mitigation: Critical risks shown in the PAD were appropriate but not framed to focus on the real elements likely to cause major difficulty. Mitigation measures were overly standardized, ratings were mostly Moderate when deeper analysis and more realism would have suggested Substantial or High, and some risks were not acknowledged: how a small, new management unit would coordinate multiple, established agencies; the depth of political rivalry in the State of Maranhão; the political economy affecting resettlement and its operationalization; project reliance on independently-financed and executed activities/works and with horizontal and vertical inter-dependencies; the integrated methodology/approach itself where unforeseen obstacles could potentially interfere with a chain of activities; and, launching such a project without a minimum set of activities ready to implement. The only risks rated as high were fiduciary. None of the design “alternatives considered” included the option of a simpler, more limited/selective, initial approach. The PAD refers briefly to the project as a “pilot” and a “phase” but project design does not reflect such or outline a longer-term framework.

3 Peer Reviewers questioned whether the project really was an integrated urban/water management operation or an urban/environmental project with many activities in different sectors. An integrated operation would more typically see a series of combined investments/activities designed to minimize costs and maximize benefits, ensuring efficiency gains.

8

2.1.11 Implementation arrangements: The UGP within SEMPE – a new agency created for the Project – was to coordinate Project implementation, working with multiple secretariats of the municipal government to ensure timely and integrated execution. 4 As shown in the PAD, SEMPE/UGP’s responsibilities were extensive; the Loan Agreement required that it be supported by a contracted firm. Opinion varies on such arrangements which can be expensive, perform poorly and be difficult to control as the administration/oversight of the contract often depends on the inexperienced project coordination unit which justified their hiring in the first place. Contracting international technical consultants at critical stages may have worked or, embedding personnel of an international technical cooperation agency in the UGP for the duration was an option with the advantage of providing mentoring and continuity. The State Government was to rehabilitate the Dam and CAEMA would execute part of the water supply and sewerage works, entailing certain risks discussed in Section 2.2. 2.1.12 Participatory processes: Dissemination and public consultations were conducted throughout the Bacanga Basin area to gauge the population’s needs, their response to proposed project activities including resettlement of families living in risk areas, and to determine priority areas for investment based on social vulnerability. In regard to environmental issues and resettlement, the PMSL met with community leaders throughout the project area, and disseminated the Executive Summary and preliminary Regional Environmental Assessment Report on its website and through correspondence with the main Federal, State and Municipal bodies concerned as well as non-governmental and community organizations. The intention was to bolster the Project’s social sustainability by promoting citizen participation throughout all project phases.

2.2 Implementation 2.2.1 Project implementation was uneven across components and activities. Component 1 (LED and Municipal Management Strengthening) was completed promptly, but Components 2 and 3 were troubled. Component 2 (Sanitation and Water Improvements) loan-financed macro-drainage works (Rio das Bicas and Coroado Canals) were completed, but most counterpart-financed interventions including water, sanitation and street paving financed via PAC were incomplete at closing and in some cases, paralyzed (Annex 2). Loan-financed activities under Component 3 (Urban and Environmental Improvements) had mixed results with Dam rehabilitation and land titling/regularization not executed, the latter stalled by the complex land-holding situation involving the Federal University of Maranhão (UFMA) and the State Government. The primary factors affecting Project execution are discussed below. 2.2.2 Complexity exceeded capacity to execute: The Project’s faced myriad technical, financial and contractual challenges throughout its execution; there were too many fragmented activities. The intended UGP support firm was not contracted and this was a mistake.5 Despite intensive Bank supervision, problems compounded. The final engineering designs (projetos executivos) were repeatedly rescinded due to quality and feasibility issues; numerous contracts expired, required amendment and/or needed additional works/financing. The engineering consortium responsible –

4 SEMPE/UGP coordinated: Caixa Economica Federal (PAC); State Government/State Secretariat of Infrastructure (SINFRA); State of Maranhão Water and Sanitation Company (CAEMA); Secretariat for Works and Public Services (SEMOSP); Secretariat for Urbanization and Housing (SEMURH); Secretariat of Child and Social Assistance (SEMCAS); Secretariat for Tourism (SETUR); Secretariat for Planning and Development (SEPLAN); and, Secretariat for Environment (SEMA), as well as the Pro-Moradia, Pro-Transportes and Minha Casa Mina Vida programs. 5 The requirement was resisted by the Bank and UGP: the initial volume of works did not justify it; such arrangements tended to work poorly; and, individual specialists were preferred. The 2013 Restructuring formally dropped this.

9

retained by the PMSL under an umbrella contract - was difficult to dismiss. Despite the repercussions of its poor technical performance on works quality/ timeliness, the firm remained an active project partner until 2011. Resettlement was especially difficult to operationalize, delayed by unforeseen land conflicts and complex legal, social, political and financial issues, burdening the UGP. Further, the PMSL tended to focus on the “hard” infrastructure works, neglecting the urbanization aspects designed to complement them and ensure the final impact. The Bank team and UGP strived to surmount the Project’s difficulties but its trajectory and physical progress remained erratic throughout, limiting its achievements. 2.2.3 Project costs increased: Project costs increased sharply once the detailed engineering designs were available after effectiveness, revealing the extent of cost under-estimation at appraisal and the need for additional, complementary works. The Borrower was obliged to capture significant additional CP funds to finance the increases. Another cost factor was unforeseen inflation in the domestic construction market resulting from over-heated investment in infrastructure nation-wide which caused the cost of civil works in the State of Maranhão to increase sharply (up 47.2% from 2007 to 2013, based on official data). A revised Procurement Plan (Aide Memoire September 2010) shows estimated total project cost 20 months after effectiveness had increased 113 percent (from US$59.4 m to US$126.6 m). The Borrower’s share rose from 40 percent to 72 percent with CP funding needs increasing from US$23.8 m at appraisal to an estimated US$91.0 m. The costs of CAEMA’s water supply and sewerage works had increased 297 percent and 148 percent, respectively. Amendment of the Loan Agreement in September 2010 increased the Bank’s share to 100 percent in all categories, a direct response to the Borrower’s CP funding situation (2.2.4). 2.2.4 Counterpart-financed works increased project complexity: The inclusion of wholly counterpart-financed interventions integral to achieving the PDO was inherently risky but the discovery of cost under-estimation added an urgent dimension to the project resource equation. The PMSL successfully leveraged additional CP funds from the Federal Government’s Growth Acceleration Program (Programa de Aceleração do Crescimento, PAC) which by 2010 had a generous budget for grants to qualifying municipalities. Reliance on CP funds, whether through PAC or other, similar instruments (Pro-Moradia and Pro-Transportes – taken by the PMSL as loans – and Minha Casa Minha Vida, grant-based as with PAC) added to Project complexity. Each program had its own processing and approval framework and schedule. The liberal availability of PAC financing may also have emboldened the PMSL to add new activities since it was effectively self-financing them and PAC funds were free (see 2.2.5). The larger issue however, is why this escalating reliance on CP funds did not signal the need for urgent reconsideration of project design. PAC funds flowed well initially but then became erratic due to the inefficiency of PAC’s main financial intermediary Caixa Econômica Federal (CEF) and the repeated need to revise its weak designs and technical packages. In the final year, the Federal Government delayed transfer of approved PAC funds as its fiscal situation worsened. At closing, loan-financed works had fared much better than counterpart-funded, most of which were unfinished and/or stalled. 2.2.5 The technical and environmental concept changed: The Project urbanization/technical concept and environmental approach were altered by the PMSL without Bank knowledge in the post-appraisal period, and adopted in final engineering designs after effectiveness. These changes were based on specific studies and included additional works: 6e.g., hydraulic landfills to create

6 The study Reavaliação e Detalhamento dos Estudos de Concepção Urbanistica – Relatório Final – Rezoneamento, Plano de Ocupação, Areas de Risco, Equipamentos Comunitários e Custos do Programa (2007) re-appraised the Project’s urbanization concept, presenting alternatives for the use, occupation and resettlement of the target population,

10

habitable land along the margins of the Bacanga Lake (an additional US$14.7 m) and road paving, doubled by the PMSL to 62 km (from US$1.6 m to US$26.0 m). Further, the environmental implications of these (and existing) interventions were inadequately analyzed or understood and Bank environmental safeguards requirements were virtually ignored. Bank task management rotated in late 2010 and safeguards specialists – who first documented these changes – started to join supervision missions. An in-depth Bank team review of the Project’s technical and environmental aspects prompted agreed mid-course corrections in the scope/design of works under Component 2 and adjustments to project designs for Component 3 to ensure the preservation/achievement of the Project’s integrated urban/environmental concept. 7 The entire urbanization package for the Rio das Bicas Canal/Salinas Sacavém area was reformulated in 2012 and, for the Sá Viana urbanization activities, hydraulic landfills were dropped. 2.2.6 Electoral turnover and political factors were influential: Party political rivalries between the State and Municipal Governments, electoral turnover and political commitment had negative effects on project execution (refer 2.1.6), most visibly in the State’s failure to rehabilitate the Dam and the PMSL’s lack of proactivity on resettlement. The UGP had five coordinators and three separate technical teams from preparation through closing.8 This phenomenon affects all projects in Brazil and is magnified in the case of municipal projects given that state and municipal elections alternate every two years and alignment of political parties is unusual. Electoral turnover of the PMSL in late 2012 saw the Bank attempt to promote a smooth transition for the Project, insisting that the soundly trained and experienced UGP technical staff be retained. Since all incoming governments have the right to appoint/select their own teams, this effort failed and the entire team left, to be replaced by a new Coordinator and young team under acute time and technical pressures to turn the Project around. More effective political mitigation strategies were/are needed. 2.2.7 Rehabilitation of the Bacanga Dam did not progress: Rehabilitation of the Bacanga Dam was impeded by a protracted legal/political struggle between the State and Municipal Governments for overall operational responsibility. The Bank had sought since effectiveness to negotiate/facilitate its transfer to the PMSL and at the time of the MTR, prospects seemed better given that the State Infrastructure Secretariat (SINFRA) had issued an official document asserting that the PMSL should operate and maintain the Dam as a municipal facility. However, no action resulted and the State retained control. After years of inactivity during which the State failed to honor its commitments to the Project, 2014 saw some movement in the dam rehabilitation studies as well as the Dam Contingency Plan and Operational Manual. Bank Management proposed an independent, expert legal and technical review, complemented by a carefully-managed conversation with the State Government. Months before project closing, the Bank delivered the resulting technical opinion to the State Governor and the PMSL, rejecting the State’s technical proposal for rehabilitation and warning of the risks of controlling tides through a single floodgate while repairs were executed. The State did not respond to this recommendation, nor did it conduct rehabilitation works on the Dam and in September 2015 the main floodgate collapsed. The Bank team collected information on this event and corrective measures taken, monitored the media, and kept the CMU and Bank Management informed, seeking guidance on potential actions. Notably,

e.g., creation of soil/land areas through hydraulic backfilling (aterros hidráulicos) in marginal areas along the Bacanga Lake to develop urban lots for resettlement housing. This study was cited in several final engineering designs. 7 See Nota Técnica: Salvaguardas Ambientais, A. Fortes, September 2011. 8 From project preparation through closing the Project saw four State Governors, three Mayors of São Luís, five project Coordinators, three turnovers in the UGP technical team and four Bank TTLs.

11

even with the new State/Municipal political alignment in 2015, signs of willingness to cooperate on the Dam yielded nothing. With the Project closing, the Dam was not a priority. 9 2.2.8 Resettlement was challenging and partial: The Project sought to resettle two distinct populations: families mapped into the flood-prone risk area under the project concept of Bacanga Lake water levels <2.0 meters, and those families affected by the canal works. Initially, all were covered by OP 4.12 Involuntary Resettlement. The first group was to be resettled in individual housing units on a 20 ha area in the Vila Embratel neighborhood on the Left Bank financed by the Federal Government’s Pro-Moradia (100% counterpart) program, while those affected by the Rio das Bicas Canal works would be resettled in Bank Loan-supported similar housing units on the Right Bank. However, the political economy of land use/allocation in São Luís was not adequately analyzed at appraisal, while the complexities of resettlement were under-estimated.10 In reality, the PMSL lacked commitment to or interest in involuntarily moving the flood-risk group from the Left Bank, indicating political and economic competition/pressures for land to house low income families. This also helps to explain the PMSL’s decision after appraisal to add the hydraulic landfills to the Project, increasing the supply of habitable land around the Lake.11 The PMSL worked with the UGP and the Bank to find alternative arrangements to the troubled Vila Embratel site, opting for a Federal Low Income Housing Program (MCMV – grant-based) complex in the Piancó area to voluntarily resettle the flood-risk group. 12 Meanwhile, indemnification and expropriation issues delayed the involuntary resettlement of the works-affected families around the canals, but good progress was finally made. See 2.4.3 and Annex 4. 2.2.9 Mid-term Review (MTR): With critical works still not started or only partially executed – including those exclusively counterpart-financed – and disbursements at US$10.52 m (29 percent of the Loan), the MTR of May 2012 (a 10-month delay) analyzed a possible Project restructuring, reviewed technical aspects of interventions on the Right and Left Banks, assessed compliance with social and environmental safeguards, reviewed fiduciary performance and designed an action plan for the remaining 18 months. The mission reaffirmed the validity of the PDO (the PAD version) concluding that its targets could still be achieved, with additional time. The MTR: (i) re-defined the scope of key activities such as urbanization and infrastructure associated with the Rio das Bicas Canal, the Sá Viana area and the Dam; (ii) designated activities to be dropped - mainly those included by the PMSL in the period after appraisal (see 2.2.5); and, (iii) clarified those works to be financed exclusively with counterpart resources. Safeguards still faced major challenges, especially resettlement and Dam restoration. The MTR mission pressed for accelerated Project implementation; boosted technical support for the over-burdened UGP; and, addressed the design of the first Satisfaction Survey (but not wider operational aspects of M&E and development impact). The primary output of the MTR was a detailed restructuring proposal. 2.2.10 Level II Restructuring: Restructuring the Project formally reflected: (i) technical and implementation adjustments and demands arising and/or redefined over the course of the Project,

9 See: Parecer Tecnico sobre o Projeto Executivo de Restauração da Barragem Bacanga, 2015. SINFRA advised the Bank ICR mission that due to lack of resources and emergency needs of the Dam, the State had abandoned the idea of total rehabilitation. Instead, the State would rehabilitate the stop-logs, strengthen the bridge structurally, and install one new floodgate. This plan was under discussion with the winning construction firm. Current status is unknown. 10 For example, the Project hoped to increase PMSL capacity to control and enforce land use to avoid new invasions and illegal sub-division of urban land. Feasibility of the Bicas Urban Park 2nd Stage was negated by invasions. On the Left Bank, missions noted intensified settlement of flood-risk areas and APPs without corrective or preventative actions. 11 Aide Memoire of April 2011 shows the UGP announcing that a resettlement housing complex for an additional 400 families would be built in a hydraulic landfill area on the Lake margin, not contemplated by the Project’s IRF. 12 The MCMV alternative coincided with the PMSL’s introduction of the “voluntary resettlement” idea, enabling the Municipality to induce targeted families to re-locate without becoming a budgetary burden. See Annex 4.

12

e.g., solutions adopted for macro-drainage works/canals, and inclusion of complementary urbanization activities on the Right Bank (integrated urban upgrading and resettlement in Sá Viana neighborhood); (ii) delayed implementation of key activities; (iii) significant body of investments financed exclusively with municipal counterpart, demanding the PMSL’s time/effort to secure alternative, additional resources from other programs, and to design, approve, contract and initiate the works; and, (iv) complexities affecting large-scale resettlement. See also 1.3, 1.6 and 1.7. 2.2.11 The Restructuring was intentionally modest (ISR #9, June 2013). The integrated, inter-dependent activities, institutions and sources of financing deterred un-bundling the Project for a more significant overhaul. Even so, project integration may have been more conceptual than actual in certain cases, e.g., water supply and sewerage interventions were already separate – financially and institutionally - and making little progress (although sewerage works had direct implications for the environmental health of the Lake). Dropping them required Bank leadership to explain the rationale to the PMSL – which is how the hydraulic landfills were dropped – even though this may have been politically unacceptable, as well as a potential reputational risk for the Bank. Importantly, the Restructuring occurred too late in the project cycle, and its processing was protracted/difficult. There was no inherent reason to wait for the MTR, given the Project’s acknowledged issues and, by not pragmatically ring-fencing a narrower package of interventions with greater chance of achievement, the Restructuring obliged the UGP (and Bank team) to continue struggling to implement troubled activities. Restructuring bought additional time but did not address the underlying issues affecting execution. The Restructuring Paper states: “Project scope adjustments fully maintain the Project’s original integrated and comprehensive technical concept”. See 2.3.3.

2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization

2.3.1 Design: The Results Framework (RF) provided the basis for M&E. It should be noted that all references to the PDO in the ICR mean the more complete version in the Loan Agreement whereas the Bank team used the PAD version (“Improve public service in the Bacanga Basin of São Luís Municipality and enhance the capacity of the city government in promoting local economic development and municipal management”). This complicates a critique of the Results Framework. The PDO was consistent with PMSL and Bank development priorities but its design was inadequate. It used a compound structure with dependent objectives and as worded, municipal capacity improvements in five “sectors” were to promote local economic development and improve quality of life (see table). The two PDO indicators (which can equally serve the third objective regarding quality of life) do not directly measure capacity improvements in the sectors mentioned in the dominant objective. The alignment below implies that this would occur through the end-product: consumers using services they lacked before the Project. Neither causal link – the macro linkage between the first and other objectives, or the objectives per se and their associated indicators - is well-articulated.

PDO PDO Indicators (PAD) PDO Indicators (Restructured) PDO (Loan Agreement): (i) To improve the Borrower’s capacity in the managerial, financial, urban, environmental and service-delivery areas with a view to: (ii) promoting local economic development; and (iii) improving the quality of life of the population living in the Bacanga River Basin. 1.Improve Borrowers capacity in the managerial, financial, urban, environmental and service-delivery areas

-Annual survey of population satisfaction with quality of public services and aggregate population satisfaction level improved 20% compared to initial survey -Increases in percentage of sewage collected and treated in targeted areas

-Increased consumer satisfaction with water, sanitation, drainage and urban services. - Increased percentage of sewage collected and treated in targeted areas.

2.Promote local economic development

-Adopt local economic development strategy and budget allocated to undertake main actions from strategy

-Develop local economic development strategy, and budget allocated to undertake main actions from the strategy

13

PDO PDO Indicators (PAD) PDO Indicators (Restructured) -Adopt competitiveness enhancement plan for one economic cluster by 2009 and implement recommended short-term actions from plan in 2010

-Develop competitiveness enhancement plan for one economic cluster by 2010 and implement recommended short-term actions from the plan in 2011

3.Improve quality of life of population in Bacanga Basin

-Project targeted areas acquire macro-drainage capacity to handle an intense storm every 25 years and micro-drainage capacity for a storm every 5 years -100% of families threatened with flooding from regular tidal movements of the Bacanga Lake in project targeted area be relocated to satisfaction of residents.

-Direct project beneficiaries, of which female 50% (Core) -Project targeted areas acquire macro-drainage capacity to handle an intense storm every 25 years and micro-drainage capacity for a storm every 5 years. -100% of families threatened with flooding from regular tidal movements of the Bacanga Lake in project-targeted area be relocated to satisfaction of residents.

2.3.2 The M&E agenda included: (i) Management Information System (MIS); (ii) three Consumer Satisfaction Surveys to solicit beneficiary perceptions on the quality of project-provided WSS services; (iii) Environmental Health Baseline linked to water quality monitoring of the Bacanga Lake; (iv) Mid-term Evaluation (MTE); and, (v) Final Evaluation, updating the MTE and supporting the Borrower Completion Report (BCR). This was an ambitious agenda for the UGP. 2.3.3 Implementation and Utilization: The Bank team and UGP based Project monitoring on the PAD version of the PDO. The fact that this version was considered simpler and adequate explains why it was not revised at Restructuring and why neither the original nor revised PDO indicators adequately capture the internal elements of the more complete Loan Agreement version (see table). The MIS established during preparation tracked basic physical progress but was unable to issue automatic financial monitoring reports (FMR). Progress reporting continued to fall short throughout and the evaluation agenda was weak; the former lacked appropriate monitoring of indicators and thus ISRs tended to post outdated values. Neither of the planned evaluation studies was done and the final Satisfaction Survey was still awaiting procurement at the time of ICR finalization. Several factors explain M&E performance: (i) while the UGP M&E professionals had experience, Bank supervision missions lacked an M&E specialist to mentor the UGP in best practice data collection, management and evaluation even though the PAD called for the Bank to be “closely involved in the M&E framework”; (ii) time pressures and the effort required to complete the project reduced M&E priority; and, (iii) major works remained incomplete at closing, limiting the potential scope of evaluation studies. In the final year, the Bank team explained the RF and worked with the UGP to identify gaps in data measurement and collection. The Project financed a baseline Satisfaction Survey, water quality monitoring in the Lake and a BCR, all of satisfactory quality. These reports supported preparation of the ICR.

2.4 Safeguard and Fiduciary Compliance

2.4.1 Safeguards: The Project was classified as Category A based on the Quality Assurance Team (QAT) recommendation (2005) and triggered: OP 4.01 Environmental Assessment; OP 4.04 Natural Habitats; OP 4.11 Physical Cultural Resources; OP 4.12 Involuntary Resettlement; and, OP 4.37 Safety of Dams. The following summarizes Safeguards compliance and outcomes with a more complete discussion in Annex 4 reflecting the Project’s Category A status.

2.4.2 OP 4.01 Environmental Assessment: Environmental Safeguards performance was judged unsatisfactory at exit. A Regional Environmental Assessment and Environmental Management Plan of good quality were prepared. While Bank supervision generally was intensive in the first two years, the Bank team – for reasons unexplained – did not include an environmental safeguards specialist in that period. The Aide Memoires and ISRs of this period did not comment

14

on Safeguards issues yet Safeguards compliance was rated Satisfactory. The UGP technical team of that period lacked an understanding of Bank safeguards requirements because they had never been briefed and this had repercussions for project designs/works. From 2011 through closing, Bank environmental safeguards oversight was stronger, well-reported and rated realistically. In the later period, environmental risks to the Project derived from the inexperience of the new UGP team and the constant delays in project activities. Environmental monitoring of works was satisfactory but technical monitoring and environmental licensing were uneven. Further, while ongoing works generally complied with OP 4.01, the UGP having acquired over time the required capacity, most interventions linked to sustainable environmental benefits were not completed, most importantly renovation of the Dam, sewerage systems, and full resettlement of targeted populations.

2.4.3 OP 4.12 Involuntary Resettlement: While the Social Safeguards Specialist assessed compliance under this Safeguard as moderately unsatisfactory, future resolution looks promising. An Involuntary Resettlement Framework (IRF) was the basis for a later Resettlement Plan based on identification of affected families, public consultation, and engineering/other studies. Analytical quality of the IRF was satisfactory. Social safeguards supervision was negligible in the first two years but strong thereafter. As discussed in 2.2.8 and Annex 4, there were two distinct populations to be resettled, initially aggregated under Involuntary Resettlement, but subsequently re-mapped into voluntary and involuntary groups. Progress was partial for the MCMV-financed voluntary resettlement of flood-risk families at Piancó: 108 families of the revised 568 total were settled by closing, another 260 will be settled by end-May, 2016 and the remainder await approval. Families affected by the canal works and to be resettled involuntarily saw good progress by closing under the Resettlement Action Plan (RAP) with 33 families either physically resettled or waiting in rent-subsidized accommodation. Active construction sites and level of completion were confirmed in both cases by the ICR mission. The Bank’s Safeguards Coordinator for Brazil and the Project Safeguards Specialist agreed with the PMSL - backed by an Action Plan and legal/budget guarantees - to supervise completion of the involuntary resettlement, and to verify completion of voluntary resettlement through PMSL progress reporting.

2.4.4 OP 4.37 Safety of Dams: A safety assessment study by a panel of experts during appraisal (2007) analyzed the Dam’s structural and operational conditions, found strong evidence of deterioration and imminent collapse, and issued recommendations. Performance/compliance under this Safeguard are assessed as highly unsatisfactory based on the Bank Safeguards Specialist’s analysis at exit. Despite all efforts by the Bank and PMSL to resolve the impasse associated with the State’s management of the Dam and legal obligation to rehabilitate it, the Project could not comply with OP 4.37, Safety of Dams. Political interference/rivalry played a major role in this outcome. Despite the social and environmental hazards presented by the Dam in its deteriorated condition, and the body of technical studies, recommendations and detailed designs available to the State, it was/is not a State Government priority and the prospects for its rehabilitation are uncertain. At the time of ICR finalization, the State had not responded to the PMSL’s recent overtures to reactivate a dialogue on Dam restoration. See 2.2.7.

2.4.5 OP 4.04 Natural Habitats: Project interventions sought: adequate use and occupancy of the Bacanga region; maintenance of an agreed water level in the lake; delineation of a permanent preservation strip (Área de Proteção Permanente, APP) of 30 meters; and, plan to promote use of those areas (vegetation, urban parks, leisure areas) by the population and prevent re-occupation. The Environmental Safeguards Specialist assessed compliance as unsatisfactory, due inter alia, to delays in resettling populations in risk areas and thus inability to establish the APP and execute planned works, lack of Borrower control of new, urban occupations in risk areas and in conservation areas around the Lake, and failure to improve the Dam, affecting water levels/quality.

15

2.4.6 OP 4.11 Physical Cultural Resources: The Project was not expected to affect areas of historical and cultural heritage in the City of São Luís as interventions were outside the Central Historic District. However, as the Project entailed construction and excavation to expand and replace infrastructure the EA included screening for known cultural property in the project area and incorporated chance finding procedures. The ICR assesses compliance as satisfactory. In practice, screening procedures for all works throughout the Project period did not encounter any known cultural property and chance finding procedures were not invoked.

2.4.7 Financial Management and audit: Financial Management (FM) performance was rated Moderately Satisfactory at closing and for most of the Project period. The UGP worked responsibly to address Bank FM concerns and hone capacity but certain issues persisted: (i) difficulty with systematic issuance of Financial Monitoring Reports (FMR) in the format required by the Bank. The FMR issue was never fully resolved and the UGP reverted to manual spreadsheets in Excel; (ii) quarterly expenditure forecasts and disbursement timetables inadequately aligned with project needs; and, (iii) the Project Operational Manual (MOP) was updated only once, indicative of overall pressure/priorities, not FM problems. The outdated MOP and manually-generated FMRs persisted through closing with the former consistently rated as an issue of internal control, which the ICR deems too harsh given that overall, moderate deficiencies in FM were not considered by the Bank sufficient to compromise the availability of information needed to monitor and manage project execution and resources. Audit performance was uniformly strong. Annual, independent audits of Project accounts produced clean, unqualified opinions for all years audited. The audit for 2015 is due by June 30, 2016.

2.4.8 Costs: Project costs by component/subcomponent differed markedly from both appraisal and the re-estimate at Restructuring. Total Project cost was US$26.64 m, about 45 percent of the appraisal estimate and 31 percent of the Restructuring estimate, as around US$59.0 m of planned investments (as at Restructuring) were not executed. The Bank’s share of total cost was US$23.44 m or 88 percent, compared to the appraisal estimate of 60 percent and updated estimate of 41.45 percent. The Borrower ultimately contributed only US$3.20 m (12 percent) compared to the appraisal estimate of 40 percent and latest estimate of 59 percent. Project implementation support cost was 23 percent of appraisal and 63 percent of the reduced final estimate. Annexes 1 and 3.

2.4.9 Disbursement: Disbursement was initially very slow but trended up in 2012-2013 due to the UGP’s greater agility/experience, a surge in project activity, stronger Bank supervision and the 100 percent Bank financing. Many factors already detailed foiled a linear progression of investments and disbursement, the inability of Component 3 to make headway being especially critical. In the final year, the Bank and Borrower agreed to improve disbursements by financing US$1.2 m of fast-disbursing equipment to support PMSL/SEMOSP’s maintenance of the canals. Cash compensation using Loan funds linked to efforts to accelerate resettlement was also intended to boost disbursement but as this was linked to the UGP’s request for a second extension of the closing date – denied by the Bank – it did not proceed. Some US$12.2 m of the Loan was cancelled.

2.4.10 Procurement: The Bank Procurement Specialist rated Project procurement performance Unsatisfactory at closing with risk rated Substantial primarily due to weak contract management. Three Procurement Post-Reviews (PPR) were conducted before closing and a final PPR in March 2016 focused on contract management documentation. The Bank team provided analysis, comment and clearance for a high proportion of all contracts processed due to the risks evident in the UGP’s limited exposure to IFI-financed procurement. The UGP procurement team received several rounds of Bank training and benefited from the continuity of key staff since 2009. Performance in the initial years was strong - the unit developed significant capacity in complex procurement - but

16

quality weakened due to acute time/related pressures. The Bank’s comments on key issues, e.g., contract cost composition, went unheeded. Some execution delays affecting large contracts in the final period were avoidable, and virtually guaranteed that the works in question would not be completed by closing. The UGP felt that the Bank’s response time on some no objections was too slow, and its interpretations unclear. However, the Bank team and Procurement Specialist strived to support the PMSL/UGP in achieving maximum physical execution of works pre-closing.

2.5 Post-completion Operation/Next Phase

2.5.1 A follow-on operation is not under consideration at this time. The Bank and Borrower signed a Management Letter (2015) establishing the measures the Borrower should take to complete all works unfinished at closing. This cannot be guaranteed given the fiscal crisis at the central level but the UGP and PMSL have the inputs to complete implementation: detailed engineering designs for all WSS and urban upgrading works ready for procurement; secure funding for WSS in the Left Bank; CAEMA’s WSS works on the Right Bank are 30 percent completed and still in the budget pipeline; and an experienced, core, follow-up team has been maintained on the UGP. On resettlement, the Bank will supervise completion of pending involuntary resettlement under formal arrangements, and monitor voluntary resettlement through PMSL notifications. Meanwhile, the PMSL continues to pay the “social rent” – at rates higher than average for such projects – to families waiting for new housing on the Right Bank (see 2.4.3). The impasse affecting the Bacanga Dam is unlikely to be resolved soon (see 2.4.4). O&M of the canals is incipient. The Bank team sent a clear message to the Municipality/UGP on the need for systems, behavior modification and organized use of the equipment purchased for canal maintenance.

3. Assessment of Outcomes

3.1 Relevance of Objectives, Design and Implementation 3.1.1 Relevance of objectives (Substantial). Taking into account the conditions on the ground in the Bacanga Basin (Sections 1.1-1.4) the PDO’s diagnosis of development priorities remained at closing, highly relevant to Bank Country Partnership Strategies (CPS, 2008-2011 and 2012-2015), to corporate goals and to the objectives of the Brazil Municipal Lending Program. The latter CPS called for: continued lending to sub-national governments; enhanced implementation of strategic social and infrastructure investment programs; improved quality and expanded provision of public services for poor households; and, increased efficiency of public (and private) investments. More directly, at the municipal/local level, the PDO remained well-aligned with the PMSL’s strategic instruments: (i) Multi-Year Development Plan (2014-2017), emphasizing among its 10 priority themes improved governance, public sector management, urban development and services delivery, and quality of life; and, (ii) Long-Term Strategic Plan – Sao Luis 2033, with its 15 priorities including local entrepreneurship and sustainable development of the Itaqui/Bacanga region, expanded access to sewerage and water supply, land regularization and housing access, the preservation and recuperation of environmentally sensitive areas, and improved public sector management. That said, and using a conservative approach, balancing the PDO’s high strategic relevance against the framing of the objectives - ambitiousness relative to institutional capacity, activities financed and engagement period - reduces its relevance from high to substantial. 3.1.2 Relevance of design and implementation (Modest): Project design sought to maximize project benefits and reflected defined local needs but its relevance to specific elements of the PDO was diminished by several factors: (i) scale and type of activities could not capture the high level of the capacity-building objective which spanned five sectors of government. It was also difficult

17

methodologically to measure such gains and to attribute them to the Project and, to determine the causal chain between actual capacity growth in those sectors and improvements in the two dependent objectives – LED and quality of life”. The investment activities were intended to leverage the implemented managerial reforms; (ii) quality of the Results Framework was uneven (even after the Restructuring), and not adequately aligned to the capacity building or LED objectives. Quality of life indicators were more directly relevant (and their related activities had high relevance); (iii) activities per se were too numerous, complex and inflexible to ensure that the objectives and intermediate outcomes would be achieved, and as noted, the overall strategy was flawed; and, (iv) the quality of the Bank’s implementation assistance was uneven (Sections 2.2 and 5.1), e.g., late Restructuring, insufficiently responsive to the situation on the ground and not using the opportunity to improve relevance of the Results Framework, limiting the Project’s ability to focus on feasible interventions and to measure their impact. See Section 2.3.

3.2 Achievement of Project Development Objectives (Efficacy) Rated: Modest 3.2.1 The following assessment breaks the PDO down into its three objectives and aligns available evidence for achievement based on the RF and other information, bearing in mind that at the time of ICR conclusion: major activities linked to the PDO were incomplete and/or stalled; no evaluation was conducted; and the Satisfaction Survey was still awaiting procurement. Objective 1: Improve the Borrower’s capacity in the managerial, financial, urban, environmental and service-delivery areas Rated: Negligible 3.2.2 Neither of the two indicators aligned to this objective captures it directly. Managerial capacity improvements are limited to assessing their implied effects, i.e. can be gauged but not substantiated. There is no direct, evaluative data on the nature and depth of such changes, which specific “area” improved and whether such improvements can be attributed to the Project. 3.2.3 Indicator 1 under PDO 1 was to: “Increase consumer satisfaction with water, sanitation, drainage and urban services”, rated Negligible. Consumer satisfaction with service delivery could not be formally assessed due to lack of evidence. Water and sanitation works were not completed and urban services only partially. However, the Rio das Bicas and Coroado macro-drainage canals were renovated to a high standard. Extensive paving works were completed along the margins of the canals, along with urban parks and plazas with communal facilities and equipment, and modern street lighting (see photos). The ICR team observed active housing renovation fostered by improved conditions around the canals, as well as small, new commercial sites and community use of the parks and open leisure/transit areas associated with the canal works. Canal maintenance needs focused attention if project benefits and consumer satisfaction are to be sustained. Scale is relevant: the Project estimated that 92,500 people would benefit comprising 77,500 from WSS and 15,000 from macro-drainage works. Actual beneficiaries of macro-drainage were 19,513, 130 percent of target, and of WSS, were zero. 3.2.4 Indicator 2 under PDO 1 was to: “Increase the percentage of sewage collected and treated in targeted areas”, rated Negligible. Works for the sewerage system on the Left Bank (PMSL CPF) were stalled at closing. The system on the Right Bank, under CAEMA, had reached 30 percent execution. Revised and complete designs for the sewage collection system financed by the Loan allowed the PMSL to keep funding for these investments secured for post-project completion. The planned sludge drying facility and pilot solid waste management activities were dropped by the Restructuring, transferred to other institutions and financed by CP resources not associated with the Project. Their status is unknown.

18

3.2.5 Other supporting information: The following institutional capacity gains are also relevant, gleaned from interviews with current and former Project coordinators and specialists: (i) deeper understanding of how to manage all critical aspects of resettlement; (ii) greater adeptness in capturing development investment resources from national and multilateral institutions/programs; (iii) increased knowledge of Bank procurement rules and sophisticated contracting processes; (iv) improved financial management of large-scale projects; and, (v) better understanding of environmental issues, safeguards compliance and environmental licensing. Objective 2: Promoting local economic development Rated: Substantial 3.2.6 Achievement under this objective is inferred from Project inputs (training, seminars and partnerships) and outputs (trained/employed workers, studies, strategies developed/adopted, action plans) designed to promote local economic development. Impacts were not measured. Similarly, the causal link between improved Borrower capacity under the first objective and promoting local economic development was not formally set up through measurable indicators. 3.2.7 Indicator 1 under PDO 2 was: “Develop local economic development strategy, and budget allocated to undertake main actions from the strategy” rated Substantial. First, the Project financed a Local Economic Development (LED) strategy which became the key analytical instrument supporting the PMSL’s short and medium-term planning instruments: the Multi-Year Plan 2014-2017, and the São Luís Strategic Plan 2033 (see 3.1.1 and archive). Second, the LED Strategy directly prompted the development of key managerial tools, for which budget was allocated: (i) portal for the Municipal Department of Economic Information and Intelligence (www.diie.com.br); (ii) Municipal Supplier Development Program to identify the suppliers of products, goods and services mainly in health, education and food; (iii) SEPLAN database with policy, decision-making and socio-economic support information; (iv) capacity-building plan for PMSL staff; and, (v) training to improve operations of the PMSL’s Permanent Procurement Center. 3.2.8 Associated with the LED Strategy, the Innovation Plan and its Action Plan13 defined four component projects: insertion of local labor into project works; identification, training and entrepreneurship guidance for micro-and small-scale local suppliers; strategic partnerships; and, environmental management. SENAI and SEBRAE trained 676 people of which 59 percent female, in basic job skills (e.g., masonry, iron-working, electrical). SEMPE fed data on persons trained under the Plano de Inovação to the construction firm responsible for the Rio das Bicas Canal works – 25 women from the trained group were hired – and established agreements with local labor unions to hire project-trained people. Courses were offered in 17 districts with 540 small enterprises participating. Environmental management courses included training for 44 youth to adopt solid waste collection and produce “eco-products” as an alternative, value-adding economic activity; and, 100 project-trained drivers are now using animal traction for local collection and treatment of solid waste. Other LED multiplier effects: (i) São Luís Economic Development Council, a new unit in SEPLAN to capture development resources; (ii) partnerships with the State Government in urban mobilization and road renovation; and, (iii) new Sub-secretariat for Sustainable Development (SADES) in SEPLAN, responsible for the overall LED agenda. 3.2.10 Indicator 2 under PDO 2 was “Develop a competitiveness-enhancement plan for one economic cluster by 2010 and implement recommended short-term actions from the plan in

13 Plano de Inovação da Bacia do Bacanga, SEBRAE, 2012; Plano de Inovação da Bacia do Bacanga: Plano de Ação, SEBRAE, 2012.

19

2011”, rated Substantial. Tourism was selected as the target cluster. Tourism Cluster Development studies 14 and Tourism Marketing projects/activities to boost tourism to São Luís and create economic opportunity, focused on: (i) selling São Luís as a “brand”; (ii) increasing foreign tourism (still low but growing); (iii) establishing a Portal to encourage industry cooperation and articulation, and a Tourism Map (hotels and restaurants); (iv) foreign benchmarking visits by tourism industry stakeholders; and, (v) agreements between the Secretariat of Tourism and airlines to promote São Luís. Finally, SEMPE is developing a US$12.0 m first-stage project with IDB to rehabilitate the Central Historic District of São Luís, a direct outcome of the above activities. Objective 3: Improving the quality of life of the population living in the Bacanga River Basin Rated: Modest 3.2.11 Assessment of quality of life improvements is based on a limited set of investments, partially concluded in the case of resettlement. No impact evidence is available. First, macro-drainage works inspected by the ICR mission and compared to original conditions showed well-constructed physical structures and surrounding areas. Direct evidence suggests this is promoting community renovation of local residential and commercial property. Beneficiaries stated that the flooding of homes and businesses had ceased since canal completion. The main caveat is canal maintenance, which is incipient and could erode benefits if not addressed. Second, resettlement of families (2.2.8, 2.4.3 and Annex 4) has delivered a higher physical standard of living. The scale of benefits is expected to increase as resettlement is completed. Third, local road paving/storm drain works and urbanization (parks and plazas) inspected by the mission are providing a normalized and more secure environment for families, with direct evidence of utilization. See Annex 11 photos. 3.2.12 Indicator 1 under PDO 3 was: “Direct project beneficiaries of which female 50 percent (Core)”, rated Modest. This indicator was intended to cover WSS and drainage only. Some 19,513 people benefited directly from macro-drainage works (130 percent of the target of 15,000) of which about 51 percent were women in the case of Rio das Bicas and 53.3 percent in the case of Coroado. The WSS works were not executed. To add, 141 families benefited from resettlement and another 5,450 from the parks and plazas, before project closing. 3.2.13 Indicator 2 under PDO 3 was: “Project-targeted areas acquire macro-drainage capacity to handle an intense storm every 25 years and micro-drainage capacity for a storm every 5 years” rated Substantial. Macro-drainage works were completed at the Rio das Bicas and Coroado Canals. Engineering designs reviewed by the UGP and Bank team showed the canals as technically adequate to handle a 25 year intensity probability storm. At ICR conclusion, they had yet to be tested under such severe conditions but had handled repeated heavy rainfall events effectively. Micro-drainage works to handle a 5-year storm were not completed but they constituted only a minor segment of the drainage intervention. 3.2.14 Indicator 3 under PDO 3 was: “100% of families threatened with flooding from regular tidal movements of the Bacanga Lake in project-targeted areas are relocated to the satisfaction of residents”, rated Modest. Some 108 flood-risk families were relocated and another 260 families will follow by end-May 2016. The Satisfaction Survey was unavailable to assess resident opinions but the ICR mission interviewed new residents and community leaders on-site, finding high levels of satisfaction with visibly improved living conditions (2.4.3, 3.2.11, Annex 4).

14 Plano de Desenvolvimento do Cluster de Turismo de São Luís, Sociedade Portuguesa de Inovação (SPI), 2011; Projetos de Desenvolvimento do Cluster de Turismo de São Luís, SPI, 2012.

20

3.3 Efficiency Rated: Negligible 3.3.1 Methodology: An economic analysis was conducted to evaluate project works implemented, using actual costs and actual benefits obtained from drainage, resettlement, and urban parks investments, i.e., interventions for which there were beneficiaries. The cost of evaluated works corresponded to 65 percent of the total cost of interventions. Annex 3 details the methodology and results of the economic evaluation compared to appraisal. Restructuring in 2013 did not update that analysis due to the completion status of works. Results tend to the conservative side, not accounting for all attained benefits. The unit cost of all interventions was estimated and compared to what was expected. Costs and benefits were converted to 2008 prices to compare with appraisal, adjusting for currency fluctuation, 2008-2015. Inflation in this period was 62 percent. 3.3.2 Results: Results for water interventions and sewerage were zero because they were not executed. Drainage results using a cost effectiveness analysis showed that actual costs were half the alternative selected at appraisal: total cost was lower and number of beneficiaries higher. Results of the cost benefit analysis conducted for drainage, urban parks, and voluntary resettlement show that the investments had positive impact on the area intervened. Actual benefits were 27 percent higher than actual costs and interventions showed a 14 percent return, higher than the 10 percent discount rate. All interventions showed benefits exceeding costs and returns of 13 percent or higher. Results are reassuring given than other important benefits could not be quantified either for lack of information or for being intangible and difficult to estimate. 3.3.3 Despite the positive results reported above, and direct evidence that interventions are improving residents’ daily lives, efficiency is rated Negligible given that: (i) the expected scope of the Project was wider and included WSS, which if implemented would have brought higher benefits; (ii) systematic maintenance is not yet in place, potentially jeopardizing sustainability and the benefits already enjoyed by residents; (iii) hard analytical evidence of impact is lacking; and, (iv) the 22 month extension did not deliver expected results and one-third of the Loan was cancelled. 3.4 Justification of Overall Outcome Rating Rating: Unsatisfactory 3.4.1 The overall outcome rating of Unsatisfactory is based on the following:

Relevance: Modest The PDO, while ambitious, reflected a well-diagnosed set of priorities aligned to past and current Bank and Borrower strategies. But, the ability of its physical design/activities to achieve those priorities – especially in the time allotted - varied among components, and the PDO was not matched by a well-designed RF which could measure achievements and substantiate the causal link.

Efficacy: Modest This rating takes into account the following and compares it to what was intended: scope, scale and level of completion; beneficiary numbers; likely sustainability of completed works; and, prospects for post-Project completion of unfinished works. Also important is the lack of hard evidence/data on impact, although indirect, anecdotal and direct visual evidence is substantial, as noted above.

Efficiency: Negligible The efficiency rating, in addition to the analysis (Annex 3), factors in sustainability, evidence, 22-month extension and cancellation of US$12.2 m of the Loan.

3.5 Overarching Themes, Other Outcomes and Impacts (a) Poverty Impacts, Gender Aspects, and Social Development

21

3.5.1 Poverty: The Project had indirect poverty reduction goals through its LED and employment-creation activities (Component 1) and direct impact by filling 25 (known) construction jobs. See 3.5.2. If poverty of assets and basic living conditions are considered, resettled families have significantly improved their housing standards while 19,513 people have had the risk of periodic flooding greatly reduced, noting the caveat regarding O&M. 3.5.2 Gender: Under the Component 1 Innovation Plan and based on the recommendations of Project-financed studies, 402 women were trained for employment in the construction industry, about 59.5 percent of the total 676 beneficiaries of training under the “Inserção de Mão de Obra” program under the LED Innovation Plan. Of these, it is known that 25 were employed on project canal works. Further, about 51 percent of macro-drainage beneficiaries were women, and 53.3 percent of resettled families so far have been female-headed households. See studies, Annex 10. 3.5.3 Social development: Local, direct community participation, public meetings and consultation had positive social effects: (i) engaged communities associated with all major works; (ii) reduced the numbers defined as potentially affected by the Rio das Bicas Canal from 180 to 33; (iii) resulted in families accepting the technically preferable open (vs closed) canal option along with enhancements such as plazas and greenery; (iv) led to project-affected communities having a seat on local councils and input into decision-making; (v) provided decisive input into the “social rent” devised by the UGP to tide families through the housing construction period, permitting the canal works to proceed; and, (vi) helped the PMSL’s negotiation with the SPU on land regularization, paving the way for future titling of occupied land in the Bacanga region. (b) Institutional Change/Strengthening 3.5.4 The Project’s impact on institutional growth and longer-term capacity was not studied formally but certain benefits are implied, primarily for the coordinating body and PMSL. The Bank insisted that the UGP be located within a secretariat but since most lacked the capacity to “host” the Project, SEMPE was established. Since project closing, its status was converted to permanent Secretariat for Special Projects and it is now developing projects with IDB and other entities. A core group of current UGP specialists remains in SEMPE or may relocate within the PMSL, with clear technical/experiential advantages. UGP team members saw institutional and personal growth in the multiple challenges of implementing and managing a large project. Informally, and based on interviews with past/present UGP coordinators/specialists, institutional gains include capacity to manage resettlement; handle the role of direct counterpart of an IFI; capture/leverage CP funds; manage environmental compliance; and, conduct large-scale, sophisticated procurement. The PMSL was resourceful in raising large amounts of development financing and was strengthened by its exposure to an IFI and a complex development operation. Procurement training for SEPLAN and CPL was important. See also 5.2. (c) Other Unintended Outcomes and Impacts (positive or negative) N/A

3.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops N/A

4. Assessment of Risk to Development Outcome Rating: Significant 4.1.1 The risk to development outcome balances positive and negative factors and is overall, rated Significant: (i) Positive: As detailed in 2.5.1, the UGP and PMSL have the inputs to complete key interventions, the main risk being the deteriorating fiscal situation in Brazil; and, (ii) Negative:

22

Three interventions have substantial risk of non-execution: the Dam, dependent on State funding; and, the Rio das Bicas Urban Park and urbanization/environmental works on the Left Bank, both Loan-financed. Drainage works conform to high design and construction standards but maintenance is still being organized. SEMOSP is responsible for canal upkeep and while it has appropriate equipment, it lacks systems and planning capacity. As the sewage outflow issue was not addressed, the canals show a mix of sewage, rainwater and vegetation growth, risking renewed flooding. Similar observations are relevant to the Piancó housing clusters where new buildings and public areas were already showing signs of weak maintenance. Community leaders interviewed by the ICR mission were actively promoting organized O&M of buildings by residents. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry Rating: Unsatisfactory 5.1.1 To avoid repetition, Section 2 provides substantial detail on preparation, design and quality at entry relevant to this discussion. The Project was well-anchored strategically, its preparation was consultative, inclusive and supported by studies (albeit of variable quality). It focused on an integrated approach, although not for the typical efficiency rationale (ff. 4). Poverty and gender concerns were reflected in the Project’s quality of life and LED goals/activities. Efforts were made to satisfy the PMSL’s desire for an operation which would resonate with stakeholders and recruit their collaboration. Agreements were signed by the PMSL with the State Government and CAEMA to ensure their adherence. Even so, preparation was prolonged and expensive, yet the Project was not ready to implement. Design far exceeded the client’s capacity and was based on flawed assumptions/analysis – especially in regard to resettlement – and under-estimated risks. The Project over-reached for a single, five-year operation; allowed activities critical to achieving the PDO to be wholly counterpart-financed; and, assumed that the State Government would collaborate in a municipal project irrespective of political alignment. The Loan was too small for the proposed activities once realistically costed, obliging the Borrower to find the considerable additional funds needed. The Bank team acquiesced in the PMSL’s drive to leverage the most impact possible from a single operation, rather than guiding the Client in controlling scope and scale. The Bank’s internal review and clearance process was insufficiently rigorous in this case. (b) Quality of Supervision Rating: Moderately Unsatisfactory 5.1.3 The Project implementation period is divided into two segments, the later showing generally superior quality but both periods with positive and negative aspects. The challenges facing supervision and how these were addressed/managed emerge from Section 2, including the extent to which supervision focused on development impact, fiduciary and safeguards aspects, and M&E. Overall, project supervision intensity/inputs and specialization exceeded Bank norms - 25 missions including fiduciary and technical – but not always staffed for project needs. Fiduciary supervision was strong throughout. However, in the first two years, supervision quality suffered from: lack of safeguards oversight; undue lenience towards the Borrower’s changes to the project concept/activities; not insisting on replacement of the under-performing engineering firm; apparent “distant” relationship with the UGP technical team; less than candid reporting/ratings; and, not contracting the management entity required under the Loan Agreement.

23

5.1.4 Superior supervision from 2011 included: proactivity on resettlement and the Dam and efforts to re-define and clarify the Project; more thorough, candid reporting/ratings; and high quality safeguards oversight. Caveats include: postponing the MTR; delayed, overly cautious Restructuring; lack of M&E mentoring; and, advising the UGP to forego the planned management support. Overall, opportunities to restructure earlier were missed, while development impact/M&E lost focus. A more strategic approach saw the Country Director and Program Leader visit São Luís in 2013 to inspect Project works, meet the new PMSL and facilitate the planned Restructuring. Specialist support to the UGP and engagement with the State on the Dam were further intensified but problems persisted and delays accumulated. The Bank team felt its input was discounted and unable to influence the course of events. However, transitional arrangements for post-closing follow-up were well-crafted, positioning the PMSL/UGP to complete key unfinished interventions. (c) Justification of Rating for Overall Bank Performance Rating: Unsatisfactory 5.1.5 The overall rating for Bank performance takes into account: (i) unsatisfactory quality at entry; (ii) intensive supervision throughout tempered by weaknesses in the initial years which had repercussions on implementation quality and pace; (iii) generally superior overall performance over the next five years – with critical caveats – intended to reverse chronic under-performance; and, (iv) more strategic approach in the final 18 months, involving Bank management and further increasing technical inputs. Given the split rating, the outcome rating of Unsatisfactory determines the overall Bank performance rating of Unsatisfactory.

5.2 Borrower Performance (a) Government Performance Rating: Unsatisfactory 5.2.1 Government (PMSL): The PMSL in power during preparation promoted the Project concept and logic vigorously during preparation to catalyze stakeholder buy-in but was ill-equipped to gauge the complexity of the emerging design. Unilateral changes to the project concept and environmental compliance increased project complexity. The ICR suggests that this was due to inexperience, the new PMSL’s misgivings about the Project’s resettlement concept and weak operational capacity. Faced with a massive financing gap once the Project’s real costs were revealed – for which the Bank preparation team bears most responsibility – the PMSL successfully accessed additional counterpart funds. Its dialogue with the Bank team was better during preparation and with the State Government was better before signing the Loan, but both became less collegial over time as delays and problems mounted and PMSL commitment to the Project waned. Even so, the PMSL agreed to ex-post Bank supervision of resettlement, and with the terms of the Bank-sponsored Management Letter calling for completion of unfinished works. (b) Implementing Agency or Agencies Performance Rating: Moderately Unsatisfactory 5.2.2 UGP: The new UGP/SEMPE was not embedded in the institutional mainstream and thus in the initial years lacked the systemic connections and “weight” needed to mobilize established agencies, especially state-level, but this stabilized over time. The UGP also had five coordinators and three separate project teams from preparation through closing. The preparation and post-effectiveness teams acquired experience and confidence only to be let go despite Bank appeals. The UGP’s tenuous relationship with the Bank team in the initial years affected its performance; Bank proactivity, collaboration and mentoring thereafter built rapport and capacity. Within these conditions, the UGP worked hard to satisfy the Bank, implement the Project effectively, and to

24

integrate and articulate the many institutions and fragmented, un-synchronized and difficult activities. The unit would have benefited from the in-house support envisaged at appraisal. Negotiations with multiple parties to resolve the large-scale resettlement of flood-prone families on the Left Bank took two years, but prompted significant maturation. The UGP felt the Bank could have tried harder to facilitate the related dialogue. Contract management and M&E remained problematic, the former due to sheer complexity, and the latter set aside under the pressure to implement the works. Even so, in the final years the UGP procured and signed new, large, contracts reflecting notable technical/administrative growth. Also, as asserted correctly by the UGP team in place since 2013, the 22-month extension was used to conclude or advance key works which had not progressed in the previous five years. The UGP’s support to the ICR was exemplary. 5.2.3 State Government/SINFRA: The State Government (SINFRA) was legally committed under the Bacanga Dam Agreement to repair the Dam floodgates and supporting structures and to operate and maintain the Dam properly (see 2.1.2). Its performance is discussed, given its critical role, but not rated as it was not formally an implementing agency (Loan Agreement Section I, A, 1.). Mostly for political reasons, the State lacked any buy-in to the Project during implementation. While the State’s role in the preparation dialogue was constructive, it was only intermittently responsive to the many overtures, negotiations and technical consultations conducted by the Bank, SEMPE/UGP and PMSL during implementation, and its follow-up was unreliable. The results of Component 1 – LED, Tourism Cluster efforts, TA/training and employment creation – could, according to former UGP specialists, have been even stronger with the State’s support. 5.2.4 CAEMA: This State agency was responsible for implementing the water and sanitation works (see 2.1.2). The institution was described at appraisal as exceptionally weak and the project experience confirmed this assessment. Neither intervention was successful due to lack of adequate planning and capacity. While water supply works were partly executed by SEMPE/UGP, no results emerged as the main water source (CAEMA’s responsibility) could not properly feed the system. Also, at Project completion the sewerage system in the Right Bank was just 30% complete. CAEMA had/still has limited technical capacity to evaluate design, plan ahead and execute works, and to operate and maintain WSS systems. As with the State/SINFRA, CAEMA is not rated. (d) Justification of Rating for Overall Borrower Performance Rating: Unsatisfactory 5.2.4 This overall rating balances the Borrower’s unsatisfactory performance with the moderately unsatisfactory performance of the UGP, against the Project’s overall outcome.

6. Lessons Learned 6.1 Lessons of Preparation:

Multiple, interrelated problems of urban poverty in Brazil require a reconsidered

strategy of phased operations consistent with the longer-term vision of states and municipalities. The level of difficulty and risk for integrated, multi-sector urban operations in fragile environments is unlikely to diminish. They need strong analytical justification, a programmatic approach with incremental implementation over a longer engagement period, focused objectives and a stable, institutional “champion”. Substantial capacity-building should accompany investment throughout, as well as a strategy for maintaining institutional and professional continuity and commitment.

Dependence on fully counterpart-financed interventions to support achievement of the PDO needs careful consideration, especially if the funds are sourced from public programs. The

25

proliferation of targeted, sector-specific financing programs in Brazil, many of them grant-based, is both an opportunity and a risk. Accessing these programs adds complexity through program-specific processing and approval criteria, schedules and financial intermediation arrangements. They also create a set of autonomous project activities which are difficult to control within the project “package”. Client cost-sharing of own development should certainly be fostered, but the potential risks and benefits of fully counterpart-financed activities should be weighed.

Ensuring a critical mass of quality engineering designs ready for procurement at effectiveness should be standard practice to identify the correct scope and scale of works, and the corresponding amount of resources and engagement period. Readiness to implement balanced with a flexible design can position a project – especially in an untried context – for more rapid take-off, moderating expectations regarding what can be achieved, building momentum and demonstrating results. Similar MLP projects in Brazil suggest the option/benefits of dividing works into well-defined/sequenced modules, adjustable depending on ambient conditions/timing.

Urban projects which include resettlement need to analyze the political and economic pressures driving land use and allocation, the risks these entail for the project and how they will be mitigated. The expropriating authority should take ownership of the resettlement process, collaborating with stakeholders to monitor/build commitment. The elements driving land allocation – scarcity, poverty, competing uses, budgets, speculation – need to be understood and reflected in project documents and design, including whether the correct/feasible approach is via involuntary or voluntary resettlement. Once land is expropriated the authority needs to ensure control over it, including by assigning social and/or public uses to prevent invasion or reoccupation. The Client’s prior agreement and tangible commitment to resettlement should be a condition of negotiations.

The political risks to project implementation, efficacy and efficiency of electoral turnover, transition and shifting political alliances in the Brazilian context are familiar but not easily mitigated. Risk mitigation needs to carefully weigh design features for their potential to add to an operation’s political challenges. Legal agreements are a legitimate instrument only if enforceable. Turnover in trained counterpart teams is unavoidable but the Bank can insist on vetting new candidates, and on the retention at minimum of fiduciary and M&E specialists. An embedded international technical agency could boost institutional capacity and continuity through such cycles. 6.2 Lessons of implementation

The decision at appraisal to provide the UGP with contracted professional support for implementation was correct, but an experienced international technical assistance agency would have been preferable to a private firm. For example, the US Bureau of Reclamation (USBR) – which performed this function for the successful Bank-supported Northeast Irrigation Project – has globally-acknowledged expertise in dams, canals, project start-up, works management and quality control, as well as O&M systems. USBR standardized documentation is now used by many Brazilian engineering firms. Embedding one or two USBR specialists in the UGP for the duration could have provided needed expertise, stability and continuity.

Restructuring can salvage a failing project but its timing and content are critical. First, the actual need for restructuring/amendment should be confronted as early as identified bearing in mind the need in Brazil for Federal Government consent. There is no inherent reason to wait for the MTR if there are compelling signs that elements of project design are losing relevance or problems are becoming chronic. Second, the internal transaction costs of changing a flawed PDO need to be balanced against the challenges facing its measurement at closing. Third, the integration of project activities should not deter comprehensive restructuring if indicated.

The Bank needs to maintain close supervision of the technical standards, compliance and performance of such projects, while establishing and institutionalizing best practices in counterpart agencies/teams. A Bank team with strong technical capabilities and expertise must be ready to intervene if technical quality and/or implementation management is threatened. More

26

specifically, should a contracted technical/engineering firm perform poorly, especially one whose work is pivotal to the project, it should be immediately required to demonstrate improvement and if incapable, should be dismissed rapidly in favor of a well-qualified substitute. The burden of this process is minor compared to the impacts on the Project of sub-standard professional performance.

A final recommendation: Multi-sector, urban and environmental management and service delivery continue to be a priority for sub-national/municipal governments in Brazil and the Bank should continue its involvement. Despite the negative results of the MLP projects – including São Luís - several considerations are relevant for future such lending. First, the Bank has an important legacy from previous urban projects in Brazil: much of the regulatory framework for WSS now in place; the “integrated” approach to urban development and water; multiplier effects seen in clients’ leveraging of additional resources/investments after the Bank-supported project; and, evidence of substantial, unaccounted-for institutional growth. Second, recent analysis of the aggregate Brazilian experience with such projects indicates: their transformational impact is longer-term; ambitious objectives must be matched by realism regarding institutional capacity and appropriate engagement period; more aggressive support is needed for institutional capacity-building, efficiency, and partnering with the private sector; and, operationally, the importance of selectivity regarding scope, scale and inclusion. Third, Brazil has a sophisticated - but not always operational - set of instruments for WSS, urban development, housing and land regularization, offering rich opportunities for Bank learning and knowledge sharing. Finally, and as further shown by São Luís, even when PAC, MCMV/other resources are available for public infrastructure investment, problems can arise in planning, evaluating and executing proposed investments, and in managing the project cycle including its coordination across levels of government. 15 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/implementing agencies 7.1 The Borrower’s letter dated April 25, 2016, responding to the Bank’s draft ICR, made the following comments (see full text, Annex 8):

The Project would have had more positive results with the benefit of an external manager (as required by the Loan Agreement) and a longer implementation period.

Despite key planned activities not being implemented, drainage, urbanization and resettlement diminished urban poverty and improved the quality of life of the beneficiary population. The performance rating for the PMSL and the Bank should therefore be Moderately Satisfactory.

Primary project indicators were built mainly around counterpart-financed works over which the UGP lacked technical responsibility and control. A project of this size should ensure that its main indicators are based primarily on Bank loan-financed activities.

Changes to the resettlement process - voluntary under the MCMV program which had its own rules and methodology – and not reducing the target, prevented its full achievement.

The PMSL remains committed to completing unfinished works, with some already concluded since Project closing, and new resources are being captured from other partners. (The letter provides updated details on the status of individual interventions).

(b) Co-financiers N/A (c) Other partners and stakeholders N/A

15 See Brazil Systematic Country Diagnostic: Realizing Brazil’s Potential and Fulfilling its Promises, World Bank, Report #101431-BR, 2016.

27

Annex 1. Project Costs and Financing

(a) Project Cost by Component (USD million equivalent) Components Appraisal

Estimate (PAD) USD m

Re-Estimate

(RP) USD m16

Actual USD m

% of RP

1. Municipal Management Strengthening and Local Economic Development

6.27 4.01 3.74 93.27

1.1 Local Economic Development and Competitiveness Strategy

0.73 2.45 2.74 111.84

1.2 Job and Income Generation Support 1.35 0.12 0.00 0.001.3 Financial Management Enhancement 0.19 0.03 0.10 333.331.4 Project Implementation Support 4.00 1.42 0.90 63.382. Sanitation and Water Improvements 34.03 48.88 16.84 34.452.1 Sewerage System 9.47 22.00 0.96 3.462.2 Storm Drain System 21.61 17.18 13.85 80.622.3 Water Supply System 1.33 8.99 1.79 19.912.4 WSS Management Strengthening and Construction Supervision

1.63 0.80 0.24 30.00

3. Urban and Environmental Improvements 15.77 32.98 6.06 28.803.1 Informal Settlement, Urban Improvement and Resettlement of Endangered Families

8.85 28.43 5.80 18.37

3.2 Bacanga Dam Rehabilitation and Lake Environmental Monitoring

5.60 0.30 0.00 0.00

3.3 Municipal Environmental Management Support

1.32 4.25 0.26 6.12

Total Baseline Cost 56.07 85.87 26.64 31.02Contingencies 3.24 0.00 0.00 0.00Total Project Costs 59.31 85.87 26.64 31.02Front-end fee IBRD 0.09 0.09 -- --Total Financing Required 59.40 85.96 26.64 31.02

(b) Financing

Source of Funds Type of Co-

Financing

Appraisal Estimate (USD m)

Re-Estimate

(RP) (USD m)

Actual (USD

m)

% of RP

Borrower Counterpart 23.76 50.73 3.20 6.30 International Bank for Reconstruction and Development

Loan 35.64 35.64 23.44 65.77

16 Re-estimated for the Restructured Project (RP, 2013). Also note that Front End Fee was not mentioned in the re-estimate.

28

(a) Project Costs and Financing by Component: PAD and Restructuring (2010)

Components PAD Restructuring (Sept 2, 2010) IBRD

(US$ m)PMSL

(US$ m)Total

(US$ m) IBRD

(US$ m) PMSL

(US$ m) Total

(US$ m) 1. Municipal Management Strengthening and Local Economic Development

3.76 2.51 6.27 3.98 0.00 3.98


0.27 0.18 0.45 2.45 0.00 2.45

1.2 Job and Income Generation Support

0.81 0.54 1.35 0.12 0.00 0.12

1.3 Financial Management Enhancement

0.28 0.19 0.47 0.03 0.00 0.03

1.4 Project Implementation Support 2.40 1.60 4.00 1.39 0.00 1.392. Sanitation and Water Improvements

20.42 13.61 34.03 19.19 30.90 50.09

2.1 Sewerage System 5.68 3.79 9.47 0.00 22.00 22.002.2 Storm Drain System 12.97 8.64 21.61 17.73 0.00 17.732.3 Water Supply System 0.80 0.53 1.33 0.00 8.90 8.902.4 WSS Management Strengthening and Construction Supervision

0.98 0.65 1.63 1.46 0.00 1.46

3. Urban and Environmental Improvements

9.46 6.31 15.77 12.47 19.83 32.30

3.1 Informal Settlement, Urban Improvement and Resettlement of Endangered Families

5.31 3.54 8.85 11.77 19.83 31.60


3.36 2.24 5.60 0.07 0.00 0.07


0.79 0.53 1.32 0.63 0.00 0.63

Total Baseline Cost: 33.64 22.43 56.07 35.64 50.73 86.37Contingencies 1.91 1.33 3.24 -- -- --Total Project Cost: 35.55 23.76 62.55 35.64 50.73 86.37Front-end Fee IBRD 0.09 -- 0.09 -- -- --Total Financing Required 35.64 23.76 59.40 35.64 50.73 86.37

29

(b) Project Costs and Financing by Component: Restructuring (2013) and Final

Components Restructuring (2013) Final IBRD

(US$ m)PMSL

(US$ m)Total

(US$ m) IBRD

(US$ m) PMSL

(US$ m) Total

(US$ m) 1. Municipal Management Strengthening and Local Economic Development

4.01 0.00 4.01 3.74 0.00 3.74


2.45 0.00 2.45 2.74 0.00 2.74

1.2 Job and Income Generation Support

0.12 0.00 0.12 0.00 0.00 0.00

1.3 Financial Management Enhancement

0.03 0.00 0.03 0.10 0.00 0.10

1.4 Project Implementation Support 1.42 0.00 1.42 0.90 0.00 0.902. Sanitation and Water Improvements

17.98 30.90 48.88 14.79 2.05 16.84

2.1 Sewerage System 0.00 22.00 22.00 0.70 0.26 0.962.2 Storm Drain System 17.18 0.00 17.18 13.85 0.00 13.852.3 Water Supply System 0.00 8.90 8.90 0.00 1.79 1.792.4 WSS Management Strengthening and Construction Supervision

0.80 0.00 0.80 0.24 0.00 0.24

3. Urban and Environmental Improvements

13.15 19.83 32.98 4.91 1.15 6.06

3.1 Informal Settlement, Urban Improvement and Resettlement of Endangered Families

8.60 19.83 28.43 4.65 1.15 5.80


0.30 0.00 0.30 0.00 0.00 0.00


4.25 0.00 4.25 0.26 0.00 0.26

Total Baseline Cost: 35.64 50.73 85.87 23.44 3.20 26.64Contingencies -- -- -- -- -- --Total Project Cost: 35.64 50.73 85.87 23.44 3.20 26.64Front-end Fee IBRD -- -- -- -- -- --Total Financing Required 35.64 50.73 85.87 23.4417 3.20 26.64

17 This is the final amount expected once the Bank is reimbursed the balance of R$305,467.83 from the Designated Account, equivalent to US$87,000 at the current exchange rate.

30

Annex 2. Outputs by Component Activities by Component and Sub-component: Original (PAD) and Additional, and Outputs by Closing

COMPONENTS, SUB-COMPONENTS

Completed (Oct 31,

2015)

Pending, Partial,

Paralyzed (Oct 31,

2015)

Not Started (Oct 31,

2015)

COMMENTS/OBSERVATIONS

Component 1: Local Economic Development and Municipal Management Strengthening (US$6.2 m, supporting the development of a Local Economic Development and Competitiveness Strategy, and TA for project execution). (a) Sub-component: Local Economic Development and Competitiveness Strategy (i) Formulation of Local Economic Development (LED) and Competitiveness Strategy, including an LED Plan (Plano DEL)

X Executed: Expected product included the planning and execution stages. (i) Socio-economic Portal for Department of Information and Economic Intelligence; (ii) Municipal Plan for Professional Qualification (conducted by Ernst and Young); (iii) Municipal Plan to Develop Government Suppliers and Purchasing (conducted by Ernst and Young)

(ii) Competitive Upgrading Strategy for critical Clusters (APLs) such as Tourism and Cultural Heritage, and a resource-based cluster focused on two major enterprises in the City of São Luís (Plano Estratégico de Marketing Turístico)

X Executed: Tourism cluster was prioritized. Activities included: development and utilization of tourism publicity images and videos; new Tourist Promotion Portal; media program and trade public relations campaigns; benchmarking trips; cooperative action campaigns/agreements with airlines; and, a program for monitoring competitors, the social reality and visibility of São Luís. The Project focused on attracting international tourism (which represent only 3% of tourists to SL). Plan is to obtain funding from the Cities Alliance/other sources. Range of other relevant activities were conducted. See 3.2.

(iii) Plan for the Development and Coordination of the implementation of the Tourism Cluster for São Luís (Plano de Desenvolvimento do Setor de Turismo)

X

Executed: Management Committee established; Tourist Map prepared and disseminated; diagnosis conducted of human resource qualifications in the tourism sector; and, tourism managers trained. Plan

31

prepared and associated Action Plan, with specific projects executed for developing the Tourism Cluster.

(iv) Innovation Plan: Development and Implementation (Plano de Inovação)

X

Executed: (i) Project for Labor Insertion: (676 people trained, 59% female); entrepreneurship training (540 small enterprises in 17 districts of São Luís). Women prioritized. See 3.2. (ii) Agreement with SINE and with the Civil Construction Union to support hiring of project-trained workers, especially women; (iii) Project to Identify, Train and provide Entrepreneurial Assistance to Potential Micro and Small Local Suppliers: (1,150 participants) (iv) Project for Strategic Partnerships: Courses offered in 17 districts of São Luís; and, (v) Project for Environmental Management was conducted under the Innovation Plan. Outputs: 44 young people trained in solid waste collection and recycling, and established a Cooperative; and, 100 drivers trained in collection/treatment of solid waste.

(v) Simplification of business processes (Doing Business) and the Municipal Balance Score Card (to be supported by IFC technical assistance).

Dropped at Restructuring

(b) Capacity-Building for Job and Income Generation Support(i) Activities such as training and facilitation of access to existing micro-finance programs

X

Executed: Training was provided for entrepreneurship targeting low income populations - including the basic steps for accessing micro-finance sources - under the Innovation Plan (see above).

(ii) TA for the Deputy Secretariat for Employment and Income Generation.

X Executed: Basic administrative training was delivered to Municipal public service staff.

c) Improved Financial Management (i) Strengthen the Municipal procurement system including assisting the CPL’s Central Bidding Committee, and fiscal information disclosure.

X

Executed: Training was delivered to the CPL and on systems for fiscal information disclosure.

(ii) Improved policy for tax collection and administration

X

Executed: Project-financed consultancy involving a diagnosis and proposal for improvement resulted in the PMSL capturing PMAT resources via the National

32

Development Bank (BNDES). Activities involved tax aspects, management of key sectors (health, education).

(iii) Improved fiscal policy Dropped at Restructuring (d) Project Implementation Support: (i) Structuring of the Project Management Unit (UGP)

X Executed: UGP structured at the time of effectiveness,

and then restructured in early 2013 following electoral change. This involved the acquisition of IT equipment and furniture; renovation of UGP premises; and, contracting of staff including technical specialists.

(ii) Strengthening of the UGP (Project Management Unit within SEMPE)

X

Executed: UGP specialists received training in Procurement, FM and Safeguards, and visited other Bank projects in Teresina/PI, São Paulo/SP, Santos/SP, and São Bernardo do Campo/SP. The Loan Agreement required that a firm be contracted to support the UGP in implementing the project. The MTR determined that this was not advisable and that the UGP was already mobilized, and receiving support from other secretariats; the Project’s Social Unit was contracted and mobilized; and the UGP had other arrangements for ensuring appropriate support. This provision was dropped from the Loan Agreement at Restructuring.

(iii) Support Project Management and Project Monitoring and Evaluation (M&E)

X Executed: The Project contracted: 4 architects; 2 engineers; 4 social assistants; 1 legal advisor; 1 economist; 1 Information Technology specialist; 1 project analyst and administrative support. Bank provided mentoring for FM, Procurement and Safeguards. Support for M&E was modest but the UGP had a permanent M&E specialist throughout.

(iv) Satisfaction Survey to assess public satisfaction with Project-financed services

X

Pending: Initial (baseline) Satisfaction Survey (SS) was completed in 2012, and a second/final SS was still awaiting procurement at the time of ICR conclusion, hampered by fiscal constraints. UGP was unable, due to acute pressure to implement the Project, to focus on conducting the planned three studies.

(v) Communication and Dissemination activities with communities

X

Executed: All project activities where relevant, were accompanied by communication, dissemination and consultation with affected groups. The Project stressed

33

participation and consultation from preparation through closing.

(vi) Audit of the Program X

Executed: Annual audits were conducted by independent audit firms and audit opinions were either Unqualified or Exception. Audit performance was strong. The Borrower/UGP followed up on any/all outstanding issues.

(vii) Training and technical visits X

Executed: UGP personnel received training at various points including in Bank Procurement, FM and Safeguards. Technical visits were organized to Bank projects in other states.

Component 2: Sanitation and Water Improvements (US$34.0 m, to improve storm water drainage, sewerage and water supply systems, and to develop and pilot solid waste management measures in targeted areas of the Bacanga Basin (BB)). (a) Sewerage System Improvements: (i) Construction of a sewerage collection system for the entire Left Bank feeding into the existing Bacanga Sewerage Treatment Plant (ETE).

X

Not executed: Works (PAC-financed) were initiated but were stalled at closing due to engineering design/other issues. Designs were being revised at the time of ICR conclusion.

(ii) Execution and completion of a wastewater collection system in targeted areas of the BB, not covered by CAEMA’s project.

X

Not executed: Works (PAC-financed) were started in late 2013 and paralyzed in 2014 due to engineering design issues. Design was inappropriate for the area planned. SEMPE contracted with AE resources, revised designs in October 2015. The investment was divided into two stages, the first valued at US$ 34.0 m. At closing, this was awaiting authorization from the Ministry of Cities to launch a new bidding process.

(iv) Municipal Integrated Basic Sanitation Plan X

Executed: Project-financed diagnostic studies were concluded. Macro-drainage project for entire Municipality to be financed by Ministry of Cities.

(v) Implementation of a mechanical sludge drying system in the ETE Bacanga.

Dropped at Restructuring

(b) Drainage System Improvements: (i) Micro and Macro-drainage for the Coroado Canal and Rio das Bicas Canal including constructing storm drain networks on the Right Bank of the Bacanga Lake/Project area.

Substantially executed: Micro-drainage (storm drain) networks, a small part of the overall drainage intervention, were not completed.

34

X

(i) Coroado Canal: Macro-drainage works concluded, showing strong community acceptance; positive social impacts observed. Environmental Safeguards in compliance. (ii) Rio das Bicas Canal: Macro-drainage works concluded, garnering strong community support; positive social impacts observed. Environmental Safeguards in compliance.

(ii) Improving the operation of the urban drainage canals by cleaning the canals, improved O&M, and removing structural obstacles in the canals.

X

Substantially executed: Canals renovated, cleared of debris and able to operate adequately. O&M practices still being established. SEMOSP is responsible but lacks systems. The Project financed heavy equipment and vehicles for canal O&M and Bank was advised after urging PMSL to start using it, that this was underway.

(iii) Rehabilitating areas of environmental value to improve the quality of water flowing to the Bacanga Lake and reduce decayed areas on the Right Bank of the Lake.

X

Partially executed: (i) Urban Green Areas Study and a Plan for the Recuperation of Degraded Areas: Both were completed but neither was executed due to shortage of time available before closing. (ii) Consolidation of the Bacanga State Park by preparing a management plan and executing recuperation activities in areas with grading issues: The Management Plan was envisaged as part of the Environmental and Social Management Plan (PGAS) but due to institutional and political difficulties, there was inadequate articulation with the State Sec. of Environment. In 2011, it was learned that SEMA had already in 2009 updated the Management Plan prepared in 1999. There is no information on rehabilitation activities. (iii) Implementation of programs to recuperate protected areas on the Lake margins (30 meter protection strip in from the 2 meter water level): Not implemented. This depended on the resettlement of the people living in those areas, and implementation of urbanization works (roads, green areas etc) on the lake

35

margins. The area continued to be invaded/re-occupied after project-resettled families left. (iv) Preservation and recuperation of environmental areas on the Left Bank: The BTT was still trying to make progress on this during the extension period for the Project but no actions were taken.

(iv) Preparation of a Plan for the Management of Solid Waste as a pilot for the Bacanga Lake.

X Not executed

(v) Review/revise the Executive Project (final design) for the Sanitation System on the Left Bank of the Bacanga Lake.

X

Executed: The final design is ready. Based on this, the Municipality can move ahead and procure the works.

(c) Water Supply System Improvements:(i) Expansion of the network/distribution system and installation of metered connections (Left Bank). (ii) Rehabilitation of distribution infrastructure linked to the Batatã Dam.

X

Not executed: (i) Not completed at Project closing date. The actual network and meters were to be counterpart-financed by UGP/SEMPE. Works for the Water Supply System in the Pressure Zone VI-A and VI-B were completed but meters were not installed due to the lack of water flow (CAEMA responsibility). Works which were under SEMOSP were transferred to CAEMA. (ii) Dropped at Restructuring.

(iv) Monitoring water quality in the Bacanga Lake and Cadmium levels.

X Executed: Quality of water in the Bacanga Lake was monitored as planned and a report prepared/available. Four collections of water were made in periods of rainfall and the dry season (Feb 14, Nov 14, April 15 and August 15) with the Bacanga Dam gates closed. Cadmium and heavy metals were not detected in relevant amounts. However, other water quality issues were detected. See Annex 4.

(v) Rehabilitation of the concrete structures and metal bridge supporting 900 mm of distribution pipes at the Bacanga Dam.

X

Executed: The water viaduct which supplies the Port of Itaqui passes through the Bacanga Dam and was in deteriorated condition with the potential to rupture. The Project financed the recuperation of the concrete structure and metal bridge in the Project’s initial years.

(vi) Loss Control Pilot Program (to be implemented by CAEMA)

X Not executed

36

(d) Technical Assistance: (i) Strengthen sanitation management and supervision of construction works, including TA to the PMSL to regulate, operate and maintain water and sanitation services.

X

Partially executed: Technical assistance/training was limited overall. Refer to item (iv) below. O&M of water and sanitation investments was dropped at Restructuring, as it constitutes an inherent responsibility of CAEMA.

(ii) Improve O&M of drainage services. X

Partially executed: However, SEMOSP acquired US$1.2 m of heavy equipment and trucks for maintenance of the canals and Bank was advised at the time of ICR finalization that this was underway.

(iii) Develop a Solid Waste Management Program.

X

Executed: Modified by the Restructuring and completed. Originally, the PAD called for a Solid Waste Management Plan and during Restructuring it was adjusted, substituted by development of an Integrated Sanitation Plan (covering water, drainage, sanitation and waste).

(iv) Strengthening of sanitation management/Plan for Director of Sanitation

X

Executed: As explained above, a Municipal Integrated Sanitation Plan was prepared and sent to the Municipal Attorney General’s Office to review the draft Law for approval by the Municipal Chamber of São Luís.

(v) Supervision of sanitation works X

Partially executed: As a contracting strategy, the UGP only contracted supervision for 8 months but contract execution had difficulties due to substitution of a key member.

Component 3: Urban and Environmental Improvements (US$15.8 m, supporting upgrading and resettlement in areas of informal settlement, rehabilitation of the Bacanga Dam, and Municipal environmental management). (a) Informal Area Upgrading and Resettlement(i) Studies for establishing legal limits of urban occupancy in the BB and to address the issue of accommodating future immigrants to the city.

X

Not executed

(ii) Plan for the Resettlement of Families

X

Partially executed: A contract prepared an Involuntary Resettlement Plan for the Left Bank defining a map of those affected by flooding, and by project works, based on a survey/cadaster of families compared to a baseline. Contracting of an engineering firm to execute the Plan did not proceed. It was not possible to construct the building at the Carneiros site

37

(Vila Embratel) and thus an alternative strategy was developed to transfer families voluntarily to the Piancó site financed by MCMV.

(iii) Financing resettlement of families in risk-prone areas (Pró-Moradia financing for Vila Embratel area, later changed to MCMV/Piancó area)

X

Partially executed: Activities/Works initiated: 108 families from flood risk and other sites in the Sá Viana neighborhood were resettled at MCMV-financed Piancó VIII apartment cluster by closing. Another 260 surveyed families were approved in March 2016 and will be resettled by end-May 2016 in Piancó I and II.

(iv) Construction of housing units for resettlement of families affected by Project works on the Right Bank

X

Partially executed:. At Project closing date, 33 housing units in Resettlement Priority Areas (APR 01 and APR 02) in the neighborhood of Sacavém in São Luís were either under construction or completed. Some 12 of the 33 families had been physically resettled, and another 21 were waiting for the delivery of their homes (scheduled for May 2016), living in rent-subsidized accommodation in the interim.

(v) Construction of local roads to facilitate access to better public services in the área – drainage and paving of 39 km (changed to paving of 62 km of existing roads).

X

Partially executed: Now stalled. Some 4.7 km was paved on Right Bank, financed by the Pro-Transporte program. This activity contemplated asphalt paving, drainage and installation of horizontal and vertical signage on the roads and avenues in the neighborhoods of Sá Viana, Pindorama, Coroadinho, Embratel and Paraíso. The contract was problematic during execution due to inconsistencies in the designs. SEMOSP revised the designs and will re-launch bidding for the works. On the Left Bank, there were no activities due to paralysis of the contract.

(vi) Hydraulic landfill, Left Bank X Dropped post-Mid-term Review. (vii) Hydraulic landfill and road paving, Right Bank X Dropped post-Mid-term Review. (viii) Construction of the Sports and Culture Plaza (Praça PEC) in the neighborhood of Coroado, and acquisition of buildings/equipment, on the Right Bank.

Partially executed: Stalled (counterpart-funded) with 70% completion. Some 8,552 m2 in 3 Pracas were implemented (Praca da Esquina 300.58 m2, Praca da Feira 515.08 m2 and Estadio de Cardosão 7,737.14 m2: green areas, playgrounds, walkways and a stadium).

38

X The large Praça PEC in the Coroado area (Right Bank), contemplated: Cultural Center with auditorium/cinema for 120; exhibition rooms; event spaces; space for courses and meetings; Reference Center for Social Assistance; covered multi-sport area; playground; cycling track; and, walking areas, in a facility of 7,000 sq meters benefiting the neighborhood of Coroado and its surrounds. The facility, which adapted a standardized, Federal Government design, remained stalled at closing through a combination of factors: unsuitable sub-soil, flow of funds, environmental licensing issues and poor contract performance.

(ix) Executive Projects, Environmental Studies and works for the Rio das Bicas Urban Park located on the Right Bank of the Bacanga Lake

X

Partially executed: The UGP approved it as a Basic Project but due to poor quality of the final design, a new firm was contracted to prepare the final design. Designs were completed, but works were only partially implemented.

(x) Executive Projects and Environmental Studies for Requalification of the Left Bank of the Bacanga Lake

X Executed: EPs, Environmental Studies and detailed engineering designs for the full Requalification of the Left Bank of the Bacanga Lake were completed.

(xi) Works for Requalification of the Left Bank of the Bacanga Lake – Lot 1 in Sao Luis.

X Partially executed: This was designed to establish the legal limits for urban occupancy in the Bacanga Basin as a way of preserving and allowing re-vegetation of areas of environmental interest. At final implementation stage, it was decided that only Lot 1 would go ahead due to lack of time. Lot 1 contemplated 4 praças plus environmental rehabilitation in Urban Green Areas (AVU). The works were contracted partially, already excluding one of the praças as the closing date was imminent. At Project closing date, Lot 1 was under conclusion with 60% completion. The AVUs were not executed. Other activities which were not part of Lot 1 included: a calçadão; a ring road de-limiting the area for rehabilitation; complementary urban parks; and the

39

rehabilitation of the Permanent Preservation Areas (APP).

(xii) Consultancy to prepare final engineering designs for urbanization, public spaces and community equipment for the Rio das Bicas Urban Park located on the Right Bank of the Bacanga Lake.

X

Executed: Final engineering designs prepared and available for procurement.

(xiii) Technical assistance to regulate land regularization for part of the population.

X Not executed

(xiv) Land regularization

X

Not executed: The regularization activity was to occur in the Left Bank only, in the neighborhoods of Sá Viana, Vila Embratel and Paraíso where the land belongs to the Federal Government (terras de marinha). However, this land was loaned to the State Government in the 1970s to build the State University of Maranhão (UEMA). This complex tenure situation (involving the SPU/Secretariat for National Patrimony, UEMA and the Municipal Government) was one of the reasons for the Project’s inability to move ahead on this activity.18 Meanwhile, the Secretariat for Urbanization and Housing (SEMURH) is conducting land regularization in partnership with the Federal University of Maranhão (UFMA) and has re-done the survey for the Left Bank of the Bacanga Lake. The Project area will also benefit from this collaboration. Also, the Project supported the PMSL in its negotiation with the SPU (Secretariat for National Patrimony) to transfer titles for informal urban areas existing around the Bacanga Basin into SPU property to promote land regularization, making it possible in a next phase to provide titles in occupied land. At the time of the project closing mission, the UGP informed the Bank

18 The area of planned regularization activity covered some 68 ha on the Left Bank in the residential zone of Sá Viana, Vila Embratel and Paraíso, comprising 2,865 residential buildings.

40

that this authorization by SPU was being finalized. This will be a positive social benefit in the Region.

(xv) Construction of public spaces and community facilities/equipment in project resettlement areas.

X

Partially executed: Urban Park Rio das Bicas: (a) Urbanization works, public spaces, community equipment in 5 praças. Originally, 9 praas were planned divided into 3 lots for bidding. One was concluded partially, another 4 were not started due to indemnification problems (land needed to be expropriated); and, (b) Final engineering designs to develop the second phase of the das Bicas Parks, including a cycling path (ciclovia) around the Permanent Preservation Areas (APP): these were procured, contracted, but not completed, due to problems with the firm.

(b) Rehabilitation of the Bacanga Dam (i) Rehabilitation of the Dam structure and rehabilitation of the electro-mechanical equipment.

X

Not executed: See Main Text Section 2.2, 2.4 and Annex 4 for history of efforts to restore the Dam.

(ii) Supervision of the Dam works X Not executed (iii) Development/preparation of operational rules for the Dam, balancing the objective of controlling floods with the environmental objective of minimizing pollution in the Lake.

X

Not executed

(c) Support for Municipal Environmental Management (i) Financing for implementing the Municipal Environmental Policy

X Not executed

(ii) Strengthening of Municipal environmental policy X Executed: A technical consultancy for development and installation, training of service providers and technology transfer to form an Information System for Environmental Management for Sao Luis (SIGA): The system was implemented and staff were trained. It is currently in the testing phase by SEMA. SIGA is designed to accelerate the environmental licensing process as well as Municipal environmental enforcement/oversight. Link: http://www.androidpit.com.br

41

(iii) Formulation and implementation of a Plan for Environmental Education for BB.

X

Not executed: The Environmental Education Plan went to bid twice, unsuccessfully. Education was to have been conducted in areas where the most serious environmental degradation and unhealthy living conditions were identified. Education was to focus on project environmental activities and social participation in this effort.

(iv) Implementation of the Plan for Managing the Bacanga State Park.

Dropped by the Restructuring

(v) Monitoring of water quality in the Bacanga Lake, and Cadmium levels.

X

Executed: Four collections conducted in rainy and dry periods (Feb 2014, Nov 2014, April 2015, August 2015) with collections made with the Bacanga Dam gates open and closed. No relevant quantity of Cadmium heavy metal was detected in the water samples. However, monitoring between August 2014 and August 2015 showed eutrophication and plankton ecology consistent with deterioration of water quality; continued pollution of the Lake with untreated sewage; and, inadequate operation of the Dam compromising the hydro-dynamic flow/flushing of the Lake.

42

Annex 3. Economic and Financial Analysis 3.1 The objective of the Project was to improve the Borrower’s capacity in the managerial, financial, urban, environmental and service-delivery areas with a view to promoting local economic development and improving the quality of life of the population living in the Bacanga River Basin. The expected beneficiaries were about 70,000 residents living in the targeted sub-basins of the Bacanga Basin. The Project consisted of three components: a) local economic development and municipal management strengthening; b) sanitation and water improvement; and c) urban and environmental improvement. In December 2013, the Project was restructured maintaining its original concept. The scope of some of the components changed: for WSS interventions the expected number of beneficiaries increased from 70,000 to 92,500 (77,500 from water supply and sanitation and 15,000 from macro-drainage works), while the number of families to resettle decreased from 750 to 595. Methodology 3.2 At appraisal, the economic evaluation for water interventions was conducted using two approaches: Cost Benefit analysis and Cost Effectiveness analysis; while for sewerage and drainage interventions, the evaluation was conducted using a Cost Effectiveness approach. Neither resettlement works nor urban parks were evaluated at that time. Results from water interventions showed expected net benefits of R$ 41.5M and return of 40.5 percent. The lifetime of the Project was estimated as 20 years. The cost effectiveness analysis of sanitation and drainage works compared different alternatives to ensure that selected investment provided the most cost effective way to achieve the environmental and flood reduction objectives. In December 2013 when restructuring the project, the economic analysis was not updated. 3.3 The fiscal analysis conducted at appraisal from 2003 to 2006 showed the PMSL’s fiscal performance was sound; the municipality’s fiscal situation complied with the Fiscal Responsibility Law. The analysis concluded that the loan from the World Bank would not place at risk this healthy financial situation. A fiscal evaluation of the municipality was not conducted for this ICR as none of the interventions implemented targeted fiscal improvement. 3.4 The economic evaluation for this ICR went beyond the one envisaged at appraisal. This evaluation was conducted for all interventions for which beneficiaries were attained, that is, macro-drainage, urban parks and resettlement. The evaluation of drainage works was conducted using two approaches: a) Cost Effectiveness, following the same methodology as at appraisal; and b) Cost Benefit analysis to complement the evaluation. The cost effectiveness analysis compared actual costs with those foreseen at appraisal as the best and more cost effective alternative. The cost benefit analysis measured the benefits as avoided damage cost when flooding was reduced. For urban parks interventions, some benefits were estimated and cost benefit analysis was used; however, additional and important benefits are intangible and were not measured. Similarly, for resettlement interventions, important benefits were not measured. Results therefore are on the conservative side and do not account for all attained benefits. To complement the analysis, the unit cost of all interventions was estimated and compared to what was expected. 3.5 In the cost benefit approach, net benefits of the interventions were estimated as the difference between the incremental benefits and the incremental cost of two scenarios: “with” and “without” the works. The “with” works scenario included actual achievements. The “without” works scenario kept the situation unchanged. The activities were appraised measuring their flow of costs and benefits for the lifetime of the works, estimated as 20 years

43

for drainage, and 30 years for resettlement and urban parks. The discount rate used was 10 percent as it is assumed was used at appraisal19. Costs 3.6 The costs of the interventions were transformed to 2008 to make them comparable with those projected at appraisal, and then transformed to economic prices using the same conversion factors applied at appraisal. Resulting costs at 2008 prices were R$ 46.6M for the whole project. See Appendix 1 of this annex for details of the approach used for the transformation of nominal prices to 2008 prices and for cost comparison of expected and actual costs.

Table 1. Actual Investment Costs per Activity Implemented (R$ m) Actual Investment Costs

Component/Subcomponent 2008 Prices R$ million

Component 1: 6.0 Component 2: Drainage Canal do Coroado 4.4 Drainage Canal das Bicas 19.6 Water Quality Monitoring 2.1 Management of Works 0.1 Sewerage and Sanitation Master Plan 0.9

Total Component 2 27.1 Component 3: Resettlement Plan 1.7 Construction of 33 houses 2.2 Urban Parks 4.5 Other interventions - urban/environmental 5.0

Total Component 3: 13.5

Total Project (without front-end-fee) o 46.6 3.7 For this ICR, the economic evaluation was conducted for: (i) drainage works along the Coroado Canal, and Rio das Bicas Canal; (ii) urban improvement through construction of recreational areas; and c) resettlement interventions. Actual cost of evaluated works corresponds to 65 percent of total cost.

Beneficiaries

3.8 Actual beneficiaries were well below expected in all subcomponents except in the drainage intervention. No works were implemented in water and sanitation. Drainage showed beneficiaries 30 percent higher than expected. Resettlement showed that some 34 percent of expected beneficiaries living in flood-risk areas had been moved, lower than predicted.

19 At time of this ICR there was no information of the discount rate used at appraisal. By appraisal the guidelines of the World Bank established that even though there was not an official discount rate to use, traditional discount rates between 10-12 percent were used to evaluate Bank-financed projects (Handbook of Economic Analysis on Investment 1998).

44

Table 2. Expected and Actual Beneficiaries

Item PAD RP Actual Actual PAD %

Actual RF %

Water, Sanitation, and Drainage 70.000 92,500 19,513 28 21 Sewerage (people) * 73,840 - - - Water (people) * 77,551 - - - Drainage (people) * 15,000 19,513 130 Drainage (households) * 4,167 5,420 130 Roads (km) 39 62 5 13 8 Resettlement (households): 20 750 595 108 0.14 18 Living in flood- risk areas 316 108 NA 34 Living in other areas 279

Economic Evaluation of Drainage Intervention 3.9 Drainage works were expected to be part of a more comprehensive plan that also included water and sanitation, which were not implemented. Actual benefits albeit lower than foreseen during preparation, have impacted positively on the livelihood of households living close to the Canal do Coroado and Canal das Bicas. As expressed by beneficiaries, the intervention eliminated persistent flooding, which not only caused damage to their residences, but also disturbed their daily lives. When flooding occurred traffic was disrupted and residents were stranded until water receded. 3.10 The construction of the canals was accompanied by urban upgrading works, such as paving, street lighting and leisure areas. Residents are enjoying not only benefits from flooding reduction, but also additional social and economic benefits: there have been active housing renovations, small business development, and residents walk safely along the paved and lighted areas. Security has improved, passive recreation is new in the area, and owners of small businesses stated that their income had doubled since the intervention. 3.11 Drainage works and urban upgrading had positive effects on the quality of life of the population living by the Canals. Residents expressed their satisfaction and how pleased they are with the works, despite the fact that a mix of untreated sewage and rainwater still runs through the Canals, due to failure to implement the expected sanitation works. 3.12 The main concern with the Canals is that maintenance is not in place yet. SEMOSP as the entity in charge of maintaining the Canals, lacks systems and planning capacity. As a consequence, the Canals are exposed to regrowth of vegetation, some dumping of garbage, and accumulation of sediments and debris at the bottom of the canals. The absence of maintenance poses a risk of renewed flooding in the area. 3.13 The evaluation of drainage works was conducted using two approaches: Cost Effectiveness to follow the same methodology as at appraisal; and Cost Benefit analysis to complement the evaluation. The cost effectiveness analysis compared actual costs with those foreseen at appraisal as the best and more cost effective alternative. Cost benefit analysis was conducted measuring the benefits as avoided damage costs when floods were reduced.

20 These are voluntary resettlement households.

45

Cost Effectiveness Analysis 3.14 At appraisal, a cost effectiveness approach was used to evaluate drainage works, and the least cost alternative was chosen. This evaluation tests actual costs of drainage works versus foreseen costs of the best alternative. The analysis was conducted using the cost per household given that the actual number of beneficiaries differed from expected. 3.15 Results shows that actual costs of the Coroado Canal and Rio das Bicas Canal were well below expected: just 47 percent of foreseen at appraisal, and 62 percent of expected at Restructuring in 2013. There is a caveat in the comparison: the scope of the works at appraisal included micro-drainage, which was not implemented. However, it is fair to conclude that even if micro-drainage were executed, actual costs would still be lower, as micro-drainage corresponded to a small fraction of total drainage cost.

Table 3. Comparison of Expected and Actual Unit Costs - Coroado and das Bicas Canals

Item PAD RP Actual (2008 prices) Total Investment Cost (US$ million) 21.6 17.7 14.38 Beneficiaries (households) 3,800 4,167 5,420 Unit cost per household (US$/hh) 5,687 4,255 2,653 Actual Unit cost/ expected Unit cost PAD 0.47 Actual Unit cost/ expected Unit cost RP 0.62

Cost Benefit Analysis 3.16 Drainage works are usually evaluated measuring the magnitude of losses and the impact the Project is expected to have in the area. Losses will vary according to intensity and duration of rainfalls, exposure, and vulnerability of assets in the intervention area. The damage may be seen on human lives, property, transport connectivity, public facilities, infrastructure, and the economy in general. Estimation of benefits is a difficult task given the stochastic nature of rainfalls and uncertainty around the vulnerability and damages of physical structures. 3.17 Among the data needed for the evaluation is the following: (i) historical data of precipitation; (ii) damage assessment of previous events; (iii) inventory of assets and vulnerability in the intervention area; (iv) flooding maps under events of different recurrence periods for two scenarios: with and without the intervention; (v) economic activity in the area and problems caused by flooding of various intensities; and, (vi) traffic count and alternate routes needed to take. For this ICR, attempts to get some information were unsuccessful and only information gathered on field visits was obtained. Households described how the flooding disturbed their daily routine and the magnitude of recurrent damages associated with it. 3.18 According to beneficiaries, the major inconvenience of flooding was: (i) traffic disruption making it necessary to take alternate routes, generated difficulties for public transportation, and increased maintenance on vehicles; (ii) people got stranded until water receded and then cleaning up was needed; and, (iii) properties were damaged. 3.19 To assess the benefits reported by households, the following approach was followed: a) main roads affected by floods were identified, though no information on traffic count was obtained and benefit from traffic connectivity could not be quantified; b) Cost of time was estimated based on information provided by households, and the minimum wage in Brazil transformed to 2008 prices (R$ 2.90/hour). Households stated that at least once a year they were stranded for one hour on average, due to flooding; after the water receded they had to

46

clean-up for about 2 to 3 hours); averted damage cost on the proprieties was based on damage costs registered at the time of appraisal/PAD (R$ 495 per household/per year) and corroborated by beneficiaries in the field. 3.20 Costs included investment and maintenance. Investment costs corresponded to the actual cost of works. Maintenance cost was included and estimated as one percent of investment. Costs were transformed to 2008 prices to make them comparable with those foreseen at appraisal. The costs were also adjusted to economic prices using the same conversion factors applied at appraisal to eliminate imperfections caused by market distortions such as subsidies and taxes21. 3.21 Results show that benefits surpassed cost by 20 percent and the internal rate of return yielded by the intervention is 13 percent. As explained before, results are on the conservative side as additional benefits such as avoided damages for more severe events, or socio and economic impacts on the area, were not included. Some 90 percent of benefits are explained by reduction of property damage, and the other 10 percent correspond to time savings. The return of 13 percent yielded by the intervention was higher than the discount rate of 10 percent.

Table 4. Results of Drainage Intervention at Canal do Coroado Summary Present Value of Cash-Flows (2008 prices)

Costs (Investment and

O&M) Benefits Net

Benefit IRR Thousand R$ 19,511 23,331 3,819 13% Thousand US$ 11,684 13,970 2,287 13%

3.22 The basic assumption for attaining these benefits is that proper maintenance is in place, which will allow benefits to be obtained during the lifetime of the Canals, estimated at 20 years. Maintenance however, has been an issue as to date it has not been addressed properly. Poor maintenance puts at risk the sustainability of the works and increases the likelihood of renewed flooding. Sensitivity analysis shows that if proper maintenance occurs, the works are viable when benefits go over fifteen years. However, if maintenance is poor and flooding is renewed, the works will not be economically viable. If floods reappear in the third year, the net loss generated will be US$ 8.4 m. If they reappear in the fourth year, net losses will be US$ 7.0 m. Proper maintenance is critical for the viability of this intervention.

Table 5. Sensitivity to period during which drainage benefits are obtained Context (US$ ‘000) IRR With proper maintenance Net Benefit 20 years 2,287 13% 15 years 405 11% With Poor Maintenance Net losses If floods reappear in the fifth year (6,024) If floods reappear in the fourth year (7,144) If floods reappear in the third year (8,377)

21 Unskilled labor: 0.46, skilled labor: 0.81, materials 0.88 and for equipment: 0.80

47

Economic Evaluation of Urban Parks and Community Facilities 3.23 The Project envisaged the following: a) 4 praças under the works for requalification of the Left Bank of Bacanga Lake plus environmental rehabilitation in Urban Green Areas (AVU), expected to benefit 15,000 people in São Viana; b) 9 praças under Urban Park Rio da Bicas, expected to benefit about 40,000 people in the neighborhoods of Pindorama, Coroadinho, and Salinas de Sacavem; and c) a large praça do esporte e cultura in the Coroado area, which contemplated a cultural center with auditorium/cinema for 120 people, exhibition rooms, event spaces, rooms for courses and meetings, reference center for social assistance, covered multi-sport area, playground, cycling track and walking area in a facility of 7,000 sq meters benefiting the population of Coroado and its surroundings. 3.24 At the time of closing 3 praças under the works for Requalification of the Left Bank of the Bacanga Lake, plus 5 praças under the Urban park of Rio das Bicas were completed or close to completion. The praça do esporte e cultura do Coroado even though almost completed at closing, remained stalled through a combination of unforeseen obstacles including unsuitable sub-soil, flow of funds, environmental licensing and poor performance by the contractor. 3.25 This evaluation was carried out only for the 8 praças completed or close to completion, whose beneficiaries are the same for the drainage works, that is, 19,513 residents or 5,420 households. Even though the Praça do Coroado may be completed in due course with State or Federal funds, it was not included in the analysis. 3.26 Facilities built in the praças include urban parks, sports amenities, children’s playgrounds, room games, green areas, a covered multi-sport stadium, and football field. Households in the area are pleased with the intervention and stated that they have, for the first time, sites for social gathering, paths to walk and do some recreation, and amenities for the kids to play and have fun. The stadium is bringing high pride to the community; several sports matches were already being planned and some sports celebrities and important teams have been invited to play there. 3.27 Parks are valuable contributors to the physical and aesthetic quality of urban neighborhoods. They positively impact public health, community building, and youth development, particularly in low income neighborhoods where facilities for recreation, sports, and leisure are scarce. Parks are valued even for those who do not use them22. 3.28 Among the benefits associated with the existence of urban parks are: property appreciation, social bonds, health improvement, and environmental impact. Many of these benefits cannot be quantified in monetary terms. For instance, value of relaxation caused by walking via footpaths is not known; same with the impact on the environment that protecting green areas brings. Other benefits can be measured through existing approaches. For instance: (i) property appreciation, can be measured through a hedonic price method; (ii) direct use of the parks and appreciation of them when not used, can be measured through willingness to pay; (iii) health impact through savings cost of medical bills when exercising, reduces doctor visits and health conditions; (iv) community cohesion, helps ward off antisocial problems that would otherwise cost the municipality more in police, fire protection, prisons, and even protecting human lives; and (v) savings on travel cost when one does not have to visit an alternative park far away from the neighborhood.

22 Godbey, Geoffrey, Alan Graefe, and Stephen W. James. 1992. “The Benefits of Local Recreation and Park Services: A Nationwide Study of the Perceptions of the American Public.” Ashburn, VA: National Recreation and Park Association.

48

3.29 Hedonic prices measure parks’ value to neighborhood quality through a statistical link between property values and proximity to neighborhood parks23. The impact that parks impose on nearby residential property values varies according to the quality of the park and its proximity. One study funded by the Graham Foundation in the US found that property value can appreciate as much as 5 percent when the property is located within 150 meters of a park24. 3.30 Studies measuring the value of parks through willingness to pay show a variety of results according to characteristics of the parks, income level of population, and other available recreational facilities close by. As example, a study carried out by the University of Passo Fundo25, in which the willingness to pay for having access to an urban park with similar characteristics to the parks executed under this project (walkways, green areas, lighting, playground) in Passo Fundo, a southern city of Brazil, found that each household was willing to pay an additional R$ 59.32 to their annual property tax. 3.31 Another approach is the travel cost method, which estimates the expenses of traveling to alternative recreational sites. In the intervention area, households stated that previously they used to go twice a month to other places to relax. Now they do not have to go that frequently as the children can play safely close to home and adults can gather and chat in the parks. They used to pay for public transportation from R$ 1.80 to R$ 2.80 per one-way trip. The distance to other sites varied widely among respondents. 3.32 This evaluation used the travel cost approach, estimating the benefits with information collected during the field visits. Additional information was expected from the results of the Satisfaction Survey that the UGP was contracting, but at time of this ICR the survey was not yet conducted. The assumptions used in this evaluation were: (i) average price per one-way trip: R$ 2.30; (ii) 50 percent of the visits to other sites were reduced; and (iii) not all members of the family went to the sites (between two and three). Cost of time spent traveling was not included as responses varied widely. Results show monthly savings on travel costs of R$ 11 per household. Benefits were transformed to 2008 prices. 3.33 Present value of benefits during the lifetime of the parks - estimated as 30 years - results in R$ 4.7 m, 62 percent higher than the costs of investment and maintenance. 3.34 Resulting net benefits from this intervention are US$ 1 m and return 18 percent; higher that the 10 percent discount rate used.

23 Phillips, Patrick. 2000. “Real Estate Impacts of Urban Parks.” Issue paper. Washington, DC: Economics Research Associates. Weicher, John C., and Robert H. Zerbst. 1973. “The Externalities of Neighborhood Parks: An Empirical Investigation.” Land Economics 49:99–105.

24 Harnik, Peter and Ben Welle. Measuring the Economic Value of a City Park System. The Graham Foundation for Advance Studies in the Fine Arts Chicago. 2009.

25 Brandli, Luciana and Pedro Prietto. Universidade de Passo Fundo. Estimating the Willingness to Pay for Improvement of an Urban Park in Southern Brazil using the Contingent Valuation Method. Article in Journal of Urban Upgrading and Development. December 2015.DOI:10.1061/ (ASCE) UP. 1943-54444.00000254.

49

Table 6. Results of Evaluation of the Praças Summary Present Value of Cash-Flows (2008 prices)


O&M) Benefits Net

Benefit IRR Thousand R$ 2,917 4,741 1,825 18% Thousand US$ 1,746 2,839 1,093 18%

3.35 These results are valid only if parks are properly maintained. If the parks are poorly managed, they can turn against the neighborhoods and instead of being relaxing places, they can be dangerous and incite violence. Economic Evaluation of Resettlement Intervention 3.36 Two actions were implemented under this intervention: (i) involuntary resettlement of families affected by drainage works; and (ii) voluntary resettlement of families residing in risk prone areas. In the first group, 33 families were resettled and awarded new housing units financed under the Project. In the second group, 108 families were resettled through an agreement with the Ministry of Cities who included them in the program Minha Casa Minha Vida, which awarded houses subsidized by the Federal Government. Involuntary Resettlement 3.37 Families affected by macro-drainage works: Project-supported macro-drainage works affected 55 families directly. Of these, the assets of 22 families were partially affected and they received cash compensation (indemnification); the assets of another 30 families were totally affected and they were offered and received new housing units26. Another three houses were not affected due to design changes but still received new housing given that it had already been offered earlier and due to their precarious living conditions. 3.38 This component was not evaluated as Cost Benefit Analysis is not appropriate to measure either the benefits or the harm that displacement can impose on the displaced families. A more detailed analysis is needed to measure actual effects on the families (positive and negative impacts), which was not available at this time. Voluntary Resettlement 3.39 The Project financed improved housing for families living in sub-standard, flood-prone settlements along the Bacanga Lake. The primary objective of voluntary resettlement was to protect families from the risks associated with the areas they were living in, and to settle them in better houses and neighborhoods in order to improve their livelihood. Some 108 families were beneficiaries under the program in coordination with the Ministry of Cities through the Minha Casa Minha Vida program. 3.40 The actual cost of the intervention was R$ 16,711 per household. Comparison to costs foreseen at appraisal and at RP was difficult as detailed information of costs included was not available. For instance, the voluntary resettlement comprises housing funded by the Ministry of the Cities under the MVMC program. It is estimated that the cost of housing financed by the

26 When design changed so did the affected area and five of these 33 houses were not affected, yet they received new housing as it had been previously offered.

50

Federal Government is about R$ 60,000 per household, which corresponds to US$ 24,000 at 2008 prices. This elevates the cost of the voluntary resettlement to US$ 33,500 per household, 63 percent of expected at RP and close to three times as much as expected at appraisal.

Table 7. Costs of Informal Resettlement Intervention

PAD RP

Actual (2008 prices)

Informal Resettlement

Involuntary (drainage

area)

Voluntary (risk prone areas) Total

Total Investment Cost (US$ million) 8.9 31.6 1.3 1.0 2.4 Beneficiaries (households) 750 595 33 108 141 Unit cost per household (US$/hh) 11,800 53,109 40,401 9,472 16,711

3.41 The evaluation of this intervention was carried out through cost benefit analysis, estimating benefits as: (i) savings on monthly rents when living in subsidized housing; this saving is partially offset by the administration fee families have to pay in the new units; and (ii) savings on damage costs associated with flooding when moving from a flood-prone area to a safe one. 3.42 Savings on rental costs were gathered in the field through interviews with beneficiaries. They used to pay a monthly rent of R$ 700 on average. In the new units they are obliged to pay only the administrative fee of R$ 53 per month. Net monthly savings are R$ 647 per household, which transformed to 2008 prices corresponds to about R$ 450. The damage cost was assumed to be the same used for the drainage intervention even though damages were higher as the area where they used to live was at higher risk of flooding and the impact higher. No better information was available. Benefits were estimated on the conservative side and actual benefits are likely to surpass this estimation. 3.43 Results show that the intervention yielded US$ 1.0 m of net benefits and 14 percent return, which is higher than the 10 percent used as the discount rate.

Table 8. Results of Evaluation of Voluntary Resettlement Summary Present Value of Cash-Flows (2008 prices)


O&M) Benefits Net

Benefit IRR R$ ‘000 4,543 6,243 1,700 14% US$ ‘000 2,720 3,738 1,018 14%

3.44 The same caveat highlighted from previous interventions has to be noted for this one as well. If no proper maintenance is given to the housing units, the benefits will be lower or it could even generate negative impact. The facilities will be declining gradually, as it was observed in the field visit, when some of the households were not paying the condominium fee of R$ 53 per month and some deterioration of the social areas of the housing complex was noticed. Community leaders however, were actively organizing residents for collective maintenance activities, building awareness and promoting a shared approach. Summary of Results 3.45 At appraisal, water interventions were evaluated using a Cost Benefit approach. The benefits were measured through consumer surplus. Results showed that expected net benefits

51

were R$ 41.0 m and expected return equal to 40 percent. There was no water intervention implemented under the Project and therefore no benefits to report.

Table 9. Comparison of Expected and Actual Results of water interventions Expected Actual

Results PAD RP Cost benefit analysis: Water Interventions

There was No evaluation

Net Benefits (R$ million) 41.5 0 Economic internal rate of return (EIRR) 40.5 0

3.46 For sewerage and drainage a cost effectiveness approach was used at appraisal. At restructuring the Project, costs and expected benefits were adjusted although the cost effectiveness analysis was not updated. There was no sewerage intervention under the Project and so no comparison can be made.

Table 10. Comparison of Expected and Actual Results of Sewerage Interventions Item PAD RP Actual Sewerage Investment cost best alternative (MUS$) 9.47 22.0 0

3.47 Results of cost effectiveness applied to drainage show that actual costs were half as much as the least cost alternative selected at appraisal. Actual costs were lower and number of beneficiaries higher. 3.48 Results of the cost benefit analysis show that the investment proved to have positive impact on the area intervened. Drainage works and urban upgrading had positive effects on the quality of life of the population living by the Canals, despite the fact that a mix of untreated sewage and rainwater was found to be running through the Canals, due to the failure to implement the expected sanitation works. 3.49 Estimated benefits from all interventions were 27 percent higher than actual costs and the return obtained of 14 percent was higher than the 10 percent used as discount rate. All interventions showed benefits higher than costs and returns higher than 13 percent in all cases. The praças showed the highest benefits, which were 62 percent higher benefits than costs and a return of 18 percent.

Table 11. Results of economic evaluation for all interventions implemented by the project

Summary Present Value of Cash-Flows (2008 prices)

(000 US$)


O&M) Benefits Net

Benefit IRR B/C Drainage 11,684 13,970 2,287 13% 1.20 Praças 1,746 2,839 1,093 18% 1.62 Informal Resettlement 2,720 3,738 1,018 14% 1.37 Total: 16,150 20,548 4,398 14%

3.50 Results are very reassuring given than other important benefits could not be quantified either for lack of information or being intangible and difficult to estimate. If interventions in water and sanitation had been implemented, benefits would have been even higher.

52

3.51 There is a significant risk that may prevent the attainment of benefits throughout the lifetime of the works, and it is related to the maintenance of the interventions. At the time of this evaluation, maintenance was incipient. If maintenance is not planned properly the infrastructure will deteriorate gradually and instead of being an asset that the community can feel proud of, it could instead turn against them and bring problems associated with violence and insecurity. 3.52 Interventions have certainly impacted positively on the daily lives of residents in the intervention area, bringing important benefits to residents, yet Efficiency is rated Negligible given that: (i) the expected scope of the Project was wider and included water and sanitation, which if implemented would have brought higher benefits; (ii) systematic maintenance is not yet in place, potentially jeopardizing the benefits already enjoyed by residents; (iii) hard evidence of impact is lacking; and, (iv) the Project was extended for 22 months but remained only partially completed, and one-third of the Loan was cancelled. 3.53 The following table synthesizes a comparison of expectations at appraisal and actual results at closing. Table 12. Comparison of appraisal expectations and actual results

Expected Actual Item PAD RP

Water Interventions

No evaluation

Approach used for Evaluation CBA Net Benefits (R$ m) 41.5 0 Economic internal rate return (EIRR) 40.5 0 Sewerage Approach used for Evaluation CEA No evaluation 0 Investment cost best alternative (US$ m) 9.47 22.0 0 Drainage

Approach used for Evaluation CEA No evaluation CEA and

CBA Investment cost (US$/household) 5,687 4,255 2,653 Net Benefit (US$ m)/EIRR 2.3/13%

Praças Approach used for Evaluation No evaluation No evaluation CBA Net Benefit (US$ m)/EIRR 1.1/18%

Informal Resettlement Approach used for Evaluation No evaluation No evaluation CBA Net Benefit (US$ m)/EIRR 1.0/14%

Total Interventions Net Benefit (US$ m)/EIRR 4.4/14%

CBA: Cost Benefit Analysis. CEA: Cost Effectiveness Analysis 3.54 See Appendix 1 for details of the approach used for the transformation of nominal prices to 2008 price and for a cost analysis/comparison of expected and actual costs.

53

Appendix 1. Actual Costs of the Project 1.1 Costs. At appraisal in 2008, the Project cost foreseen was US$ 59.0 million. During implementation, this cost proved to be inaccurate due to factors such as: (i) increase in construction costs due to the over-heated national construction market; (ii) higher costs resulting from the more detailed engineering designs done after effectiveness; and (iii) implementation of some of the interventions presenting complexities not expected at appraisal. In December 2013, the Project was restructured, and the revised cost was estimated at US$ 86.0 million. To accommodate the new financial requirements, the PMSL increased the expected counterpart funding to finance the difference of US$ 27.0 million. Costs of the interventions were adjusted and resources reallocated. Some of the interventions increased project scope, as in the case of WSS and drainage improvement which went from 70,000 direct beneficiaries expected at appraisal to 92,500. Other interventions decreased project scope, as in the case of the number of families to be resettled in the Left Bank of the Bacanga River, which went from 750 at appraisal to 595 at the time of Restructuring. 1.2 The comparison between foreseen and actual costs shows significant differences. The total cost at the time of restructuring increased by 46%. However, reality showed a different situation: actual cost was just 42% of what was predicted at appraisal, and 29% of cost expected at the time of the Restructuring. The major differences were seen with the CP funds. While the loan was 62 percent disbursed and implemented, CP funding was just 13 percent of what was predicted at appraisal and 6 percent of the estimate at the time of the Restructuring.

Table 1: Foreseen and Actual Investment Costs per Source of Funding (US $ m) Item PAD

2008 RP

2013 Actual

Total Project Costs 59.3 86.5 25.4

Differences: Actual cost/PAD 1.46 0.42 Actual cost/ RP 0.29 WB Loan 35.6 35.6 22.2

Differences: Actual cost/PAD 1.0 0.62 Actual cost/ RF 0.62 Counterpart Funding 23.7 50.9 3.20

Differences: Actual cost/PAD 2.15 0.13 Actual cost/ RP 0.06

1.3 Implementation varied widely among components. Activities implemented under Component 1 had actual cost equal to 89 percent of what was estimated at restructuring. Activities under Component 2 showed actual costs of only 32 percent of expected at RP, and Component 3 just 16 percent. The reasons for the disparities in costs are mostly found in performance rather than cost differences. Works expected in some of the components were either not executed at all or their implementation was virtually negligible, as happened with water, sewerage, and rehabilitation of the Bacanga Dam. Resettlement was also partial, micro-drainage was not done (although it represented only a small portion of overall drainage works), and other key works including urban parks and road paving were not realized or only partially.

54

Table 2: Foreseen and Actual Investment Costs per Component

US$ m Compar.

ITEM PAD 2008

RP 2013

Actual Actual PAD

ActualRP

2013 1. Municipal Management Strengthening & Local Development 6.3 4.0 3.5 0.55 0.89 2. Sanitation and Water Improvement 2.1 Sewerage System 9.5 22.0 0.76 0.08 0.03 2.2 Storm Drain System 21.6 17.7 13.6 0.63 0.76 2.3 Water Supply System 1.3 8.9 0.14 0.11 0.02 2.4 WSS management & supervision 1.6 1.5 1.79 1.12 1.19 Subtotal Component 2 34.0 50.1 16.2 0.47 0.32 3. Urban and Environmental Improvement 3.1 Informal settlement/ urban improvement 8.9 31.6 4.3 0.49 0.14 3.2 Bacanga Dam rehabilitation & Lake environmental monitoring 5.6 0.1 - - - 3.3 Municipal environmental management 1.3 0.6 1.3 1.00 2.17 Subtotal Component 3 15.8 32.3 5.6 0.36 0.17 Total Base Costs 56.1 86.4 25.3 0.45 0.29 Contingencies 3.2 0.1 - - Front end fee - - 0.09 - - Total Project Costs 59.3 86.5 25.4 0.43 0.29

1.4 The costs used for this evaluation were transformed to 2008 prices to make them comparable with those used at appraisal. To do this, the cost of the activities was broken down per currency and date of occurrence. This is important as the Brazilian currency fluctuated widely during the implementation period: the exchange rate varied against the US dollar, and also the inflation rate. Fluctuation of both exchange rate and inflation impacted on the Project in different ways: (i) Depreciation of the Real to US dollar helped the project as the Municipality received more Reais from the amount of loan funds disbursed; while appreciation of the Real did the contrary, it was disadvantageous for the project as the municipality received less Reais per amount of US dollars disbursed; and (ii) Inflation pushed costs higher, negatively affecting the Project. Disbursements from the loan were transformed to 2008 Reais using the exchange rate at the time of disbursement and then adjusted by inflation. Counterpart funds were transformed to 2008 Reais using the CPI index from the time of appraisal to the date when funds were used. Net impact of currency fluctuation on the loan was estimated comparing total disbursements actually received against the result if no fluctuation had occurred. 1.5 From the time of appraisal in 2008 to the end of the implementation period (beginning of 2016), the exchange rate went from R$ 1.67: USD 1.00 to R$ 4.20: USD 1.00, meaning a 150 percent loss of the Brazilian currency against the USD. The inflation rate in the same period was 62 percent. Each disbursement and payment was affected differently according to the time of its occurrence.

55

Figure 1. Actual costs of works, exchange rate and inflation during implementation period

1.6 Results of the cost analysis show that the average exchange rate applied to total disbursements was R$ 2.37: USD 1.00, which compared to the exchange rate of R$1.67: USD 1.00 at appraisal meant a 42 percent average depreciation of the currency against the US dollar. In other words, this meant that 42 percent more Reais were received when compared with the situation under which the exchange rate remained constant. At the same time, payments during the period were affected by 22 percent average inflation, which meant that a lower amount of goods were obtained. Depreciation benefited the Project as more Reais were available, while inflation affected negatively the Project as the amount of Reais could buy fewer goods. The net effect for the Municipality was a financial gain of 10 percent, reflected in the lower amount of USD needed27. If the currency had remained constant, the amount of USD needed to pay for the interventions would be 10 percent higher (US$ 28.0 m) than what was actually disbursed (US$ 25.3 m without front-end-fee).

Table 3: Impact of Currency Fluctuation on the Cost of the Project

Investment Cost Exchange

rate Differenc

e US$ 000 R$ 000 R$:USD % Actual Investment cost ( Nominal price) 25,322 1 59,887 2.37

Exchange rate at appraisal 1.67 42% Actual Investment cost ( 2008 prices) 46,613 -22% Benefits or (losses) due to currency fluctuation

Increase of amount received in R$ due to currency depreciation to USD

42%

Decrease of value due to inflation -22% Net benefit from currency fluctuation 10% Amount of the loan that would have

been required under constant currency 27,922

10% 1 Does not include Front-end-fee

27 Fluctuations of the exchange rate were not measured in the amount to be paid for the loan.

1.00 1.06 1.10 1.17 1.25 1.32 1.40 1.49 1.62 1.67

2.01 1.84 1.74

2.04 2.25

2.45

3.84

4.20

‐

2,000

4,000

6,000

8,000

10,000

12,000

14,000

16,000

18,000

20,000

‐

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

2,008 2,009 2,010 2,011 2,012 2,013 2,014 2,015 2,016

(000

R$)

Exchange

rate and inflation index 2008=1.00

Actual cost of works (nominal prices), exchange rate, and CPI Index 2008‐2016

Investment cost Cumulative Inflation Exchange rate R$/US$

56

1.7 To conduct the economic evaluation for this ICR, the costs were sorted per activity implemented. Specifically, drainage works along the Coroado Canal, and the Rio das Bicas Canal; urban improvement through construction of recreational areas; and, resettlement interventions, were studied separately and the corresponding costs were obtained.

Table 4: Actual Investment Costs per Activity Implemented

Component/Subcomponent Actual Investment Costs

Nominal Prices 2008 Prices Million US$ Million R$ Million R$

Component 1 3.5 7.7 6.0 Component 2: Drainage Canal do Coroado 3.3 5.6 4.4 Drainage Canal das Bicas 10.9 25.4 19.6 Water Quality Monitoring 0.1 2.6 2.1 Management of works 1.8 0.2 0.1 Sewerage and Sanitation Master Plan 0.7 1.2 0.9

Total Component 2 16.2 34.9 27.1 Component 3: Resettlement Plan 1.0 2.2 1.7 Construction of 33 houses 0.8 2.9 2.2 Urban Parks 2.5 5.8 4.5 Other interventions urban/environmental 1.3 6.4 5.0

Total Component 3 5.6 17.3 13.5

Total Project (without front-end-fee) 25.3 59.9 46.6 1.8 Comparison of foreseen costs to actual costs of interventions to evaluate, shows that drainage costs were 76 percent of expected at restructuring; and costs of informal settlement and urban improvement were just 14 percent of expected at RP.

Table 5: Foreseen Costs compared with Actual costs of Interventions Evaluated. Component/Subcomponent

Investment Costs (Million US$) Comparison PAD

RP Actual Actual/PAD

Actual/RP

Component 1 6.3 3.9 3.5 0.55 0.89 Component 2:

Total Drainage Works 21.6 17.7 13.6 0.63 0.76 Others interventions in WSS 12.4 32.4 2.6

Total Component 2 34.0 50.1 16.2 Component 3 Resettlement Plan 1.0 Construction of 33 houses 0.8 Urban parks 2.5 Informal settlement Urban Improvement 8.9 31.6 4.3 0.49 0.14 Other interventions urban/environmental 6.9 0.7 1.3

Total Component 3 15.8 32.3 5.6 Total Base Costs 56.1 86.3 25.3 0.45 0.29 Contingencies 3.2 0.1

Total Project (without front-end-fee) 59.3 86.3 25.3 0.43 0.29

57

Annex 4: Safeguards Compliance and Outcomes 4.1 Safeguards: The Project was classified as Category A based on the Quality Assurance Team (QAT) recommendation (2005) and triggered: OP 4.01 Environmental Assessment; OP 4.04 Natural Habitats; OP 4.11 Physical Cultural Resources; OP 4.12 Involuntary Resettlement; and, OP 4.37 Safety of Dams. The following presents a more detailed discussion of Safeguards-related events, compliance and outcomes (summarized in the Main Text Section 2.4.28 4.2 Environmental Assessment – OP 4.01: The Project sought to improve flood control and water quality of the Bacanga reservoir and its effluent, strengthen urban and environmental management and planning, consolidate a 30 meter permanent preservation strip around the reservoir (after re-locating resident populations under the project’s resettlement framework), consolidate the Bacanga State Park and improve solid waste management/reduction in rivers and the Bacanga Lake. The project’s Category A status required a Regional Environmental Assessment (REA) which was conducted during preparation to identify the Project’s potential environmental impacts, publicly disclosed and consulted, and was to be updated once final engineering designs became available after effectiveness. Mitigating measures for negative impacts were then included in an Environmental and Social Management Plan (EMP) with associated costs, implementation responsibilities and timetables. The REA also required that each engineering design have a specific environmental study intended to support the environmental licensing process under existing State legislation. These individual studies were to be submitted to the Bank for review/approval. 4.3 An Environmental Management Plan (EMP) covered a set of activities to prevent, minimize or compensate for environmental and social impacts of proposed interventions. These included: (i) updating the REA; (ii) Environmental Manual for project works; (iii) Environmental Education Plan for the Bacanga Basin in areas of high environmental degradation and unhealthy living conditions; (iv) implementation of a management plan for the Bacanga State Park affected by uncontrolled land use/occupation and land degradation; (v) implementation of the Municipal Environmental Policy including structuring a Municipal Environmental Licensing System and an Information System for Environmental Management of the Municipality of Sao Luis (SIGA); and, (vi) water quality monitoring in the Bacanga Lake. As noted above, final engineering designs for individual works were required to have an updated environmental assessment in each case since the designs were only conceptual/basic at appraisal, and be submitted to the State’s environmental licensing processes. 4.4 The exit rating assigned by the project Environmental Safeguards Specialist is Unsatisfactory.29 While Bank supervision was generally intensive in the first two years, the BTT and missions did not include an environmental specialist until 2011. The Aide Memoires and ISRs of this period do not comment on environmental Safeguards issues but Safeguards compliance was routinely rated in the Satisfactory range. Given the marked environmental bias/emphasis of the solutions proposed by the Project, it needed strong environmental coordination in its initial years and strong technical and environmental oversight by the Bank. Due to the lack of environmental safeguards expertise on the Bank task team, the UGP technical team of the period did not understand Bank safeguards requirements or the need to ensure that detailed engineering designs prepared after effectiveness integrated the results of updated environmental studies and that the analysis for environmental licensing purposes met Bank

28 While the ICR Guidelines do not call for formal rating of Safeguards performance, this is routinely done in supervision ISRs based on the ratings of project Safeguards specialists following each mission. In this case also, Safeguards specialists rated overall performance at exit in their final inputs for the ICR and this annex reflects those ratings. 29 See: Projeto Bacia do Bacanga: ICR – Relatorio Final Salvaguardas Ambientais, A. Fortes, March 2016

58

standards. This situation developed out of the Borrower’s important conceptual and environmental changes to the Project introduced after appraisal and with implementation initiated after effectiveness. These included additional works not contemplated at appraisal, e.g., the hydraulic landfills, which substantially changed the vision for relocating families from flood risk areas and re-developing those areas in an environmentally appropriate way, and, failure to update the Regional Environmental Assessment as agreed at appraisal or prepare environmental studies for each engineering design. Environmental studies for licensing purposes were found to be weak and did not take Safeguards into account. 4.5 The Bank Environmental Safeguards Specialist made a detailed analysis of these changes in late 2011 and in the case of environmental safeguards compliance, the Bank opted for individual technical/environmental review of each intervention – some of which had already started physical implementation – and bringing them into compliance with Bank safeguards standards. The main focus was the Right Bank macro-drainage works with environmental and urbanization aspects, urbanization on the Left Bank, designs for rehabilitation of the Dam and small adjustments to the sewerage works. The Operational Manual was also updated with socio-environmental safeguards procedures. See Main Text, 2.2.5.30 4.6 From 2011 through closing, Bank environmental oversight was stronger, well-reported and rated realistically. In the final years of the Project, environmental risks to the Project came from the inexperience of the new UGP technical team and the immense complexity of project activities which caused constant delays. While environmental monitoring was satisfactory, technical monitoring of Project interventions and environmental licensing was uneven. Further, while ongoing works generally complied with safeguards policies/procedures - the UGP having acquired the capacity to manage this process – other than the two canals, most of the critical interventions leading to sustainable environmental benefits were not completed. Under the immense pressure to execute, most planned environmentally-aligned activities had little priority or were dropped including: creation of a permanent preservation area (APP) around the margin of the Bacanga Lake (negated by the delayed relocation of eligible families in flood-risk areas as well as the failure of the PMSL to control re-occupation of the areas cleared); solid waste management improvements – sewerage and other related works – which did not progress; and consolidation of the Bacanga State Park, also affected by uncontrolled land occupation and degradation.31

4.7 In regard to the environmental impact of failure to rehabilitate the Bacanga Dam, the inability of tidal flows to flush the Lake and renovate water quality due to the rock barrier/dike installed as an emergency measure when the main floodgate failed in September 2015, created an environmental situation potentially worse than conditions at project outset. The Project’s unsuccessful efforts to implement sewerage works saw untreated sewage continuing to be discharged into the Lake and this is evident from the Bacanga Lake water quality monitoring reports, the last of which was conducted in August 2015. The report showed, inter alia, eutrophication and plankton ecology consistent with deteriorating water quality and inadequate operation of the Dam compromising the hydro-dynamic flows/flushing of the Lake waters. The Project did however, finance an Integrated Municipal Basic Sanitation Plan which was awaiting approval by the Municipal Attorney General at closing.

30 See Projeto Bacia do Bacanga: Nota Tecnica Salvaguardas Ambientais, A. Fortes, September 2011.

31 In regard to the Dam, water monitoring in August 2014 and 2015 showed: (i) eutrophication and the ecology of plankton identified in the Lake confirmed the poor quality of the water; (ii) untreated sewage; and (iii) inadequate operational conditions of the floodgates, compromising hydraulic flow into and out of the estuary.

59

4.8 In regard to compliance with the provisions of the Environmental and Social Management Plan, the following shows the status at closing:

Programs Situation at Closing 1. Environmental Management and Updating of the Regional Environmental Assessment (REA)

Up to 2011, not included in the Operational Manual (OM) and not known/understood by the UGP. From late 2011, incorporated in the OM, and satisfactory adoption of socio-environmental procedures. REA not updated until 2014.

2. Social communication Localized but proactive outreach and activities 3. Environmental Education Program still under development in 2015 4. Environmental control of project works Adequately implemented by UGP once Bank

environmental safeguards standards were understood and systems established.

5. Institutional strengthening Planned as: human resource training; agenda for environmental compliance i.e., environmental licensing and enforcement; increased material resources. Limited execution. Information System for Environmental Management (SIG) implemented in 2014/15.

6. Solid Waste Management Plan Included in the Municipal Basic Sanitation Plan developed by the Project

7.Research/investigation of sources of Cadmium pollution

Inserted into the Water Quality Monitoring – see Item 8

8. Water Quality Monitoring – Bacanga Lake and Estuary

Executed in the 2014/2015. No significant Cadmium levels detected but water quality showed specific evidence of deterioration.

9. Management Plan for the Bacanga State Park Previous plan was updated in 2009 by the State. No information on the status of its execution.

10. Resettlement of families See OP 4.12 below. Satisfactory quality IRP, partially executed with guarantees for completion in 2016 under Bank supervision. Families voluntarily resettled also partial with same guarantee of Bank verification of completion status.

Source: Fortes, 2016

4.9 OP 4.12 Involuntary Resettlement: The “requalification” or rehabilitation of flood risk areas, as well as drainage works, sanitation and re-paving of the road system around the Lake implied the resettlement of populations living in flood risk areas and those with hydromorphic soils. To ensure these aspects/features, an Involuntary Resettlement Framework (IRF) was formulated during Project preparation to guide the final designs for project resettlement and upon which a later Resettlement Plan would be based following identification of the families, public consultation, and preparation of engineering and other studies. As a Category A Project, preparation of its IRF was supervised up to appraisal by the Regional Safeguards Advisory Team, and the IRF was considered by the SAT as being of sufficient standard/quality for Board approval. All basic principles of OP 4.12 were properly covered. The Involuntary Resettlement Plan was also of good quality and sought to minimize the number of re-locations necessary in order to preserve social, economic and cultural networks existing in families’ areas of origin and to provide the services essential for quality of life in the areas to which they would re-locate. However, while the resettlement instruments were satisfactory, execution capacity was weak and operationalization of resettlement proved difficult.

60

Factors affecting resettlement:

4.10 It is important when reviewing the Project history to separate the unsuccessful efforts to implement completely Component 3 (resettlement and urbanization/environmental improvements) from the factors which explicitly affected resettlement (capacity, incentives and delays). It was initially assumed that all Project resettlement would be treated under OP 4.12 Involuntary Resettlement even though there were two distinct populations: those who would benefit from leaving flood-prone areas around the Bacanga Lake and, a group of families adversely affected by the Rio das Bicas Canal works. The 20 ha Vila Embratel/Sa Viana (Sitio Carneiros) area, originally intended for housing to resettle the flood-risk cohort, became embroiled in legal and social struggles over land ownership claims by pseudo-ethnic groups, invasions of the earmarked area, re-settlement of cleared risk areas, community leaders promoting speculation in lands they did not own, and complex issues associated with expropriation and indemnification. 4.11 An underlying theme throughout was the PMSL’s lack of commitment to or interest in involuntarily resettling the flood-risk cohort on the Left Bank, reflecting the realities of political and economic competition/pressures for land to house low income families, especially in the area of influence of two large industrial concerns in São Luís – Companhia Vale do Rio Doce and the Porto de Itaqui. Evidence suggests that immigration drivers and economic gains, i.e., the attraction to poor people of living near two major sources of employment, reduced the PMSL’s willingness to empty out the Vila Embratel area of the Bacanga River as it was a reserve of cheap land available to host poor families coming into São Luís. This rationale also helps to explain the PMSL’s decision after appraisal to add the hydraulic landfills to the Project as a means of increasing the supply of habitable land around the Lake for new settlement.32 4.12 The PMSL worked closely with the UGP and the Bank to find alternative arrangements to Vila Embratel/Sitio Carneiros, persuading affected populations – based on the UGP’s intensive community mobilization efforts and a new survey - to relocate voluntarily to a Federal Low Income Housing Program (MCMV) complex in the Piancó area which had the advantages of being situated in the same general area as Sitio Carneiros and closer to beneficiary families’ region of origin, and with construction of buildings (multi-floor apartments) already underway.33 To clarify, Involuntary resettlement is only applicable in two cases: re-taking of land by the exercise of eminent domain (the state’s power to expropriate), or, restricting access to and use of natural resources. Neither of these two factors pertained to the Left Bank cohort and thus the PMSL’s revised strategy was appropriate. The voluntary approach also reduced the contingent of families needing in the future to be involuntarily resettled if the PMSL decided to re-take land in these areas. The MCMV/Piancó solution reduced the number of families who would be negatively affected by future expropriation. Once the factors prompting activation of OP 4.12 were removed, compliance was no longer an issue for the Left Bank. 34

32 The Aide Memoire of April 2011 shows the UGP announcing that a resettlement housing complex for an additional 400 families would be built in a hydraulic landfill area on the Lake margin not contemplated by the Project’s Involuntary Resettlement Framework. This did not proceed and as stated, hydraulic landfills were dropped from the Project at Bank insistence. 33 After finding out (belatedly) that Vila Embratel/Sitio Carneiros would not work, MCMV became the late solution for resettlement, coinciding with the PMSL’s introduction of the “voluntary resettlement” approach. The MCMV alternative provided the Municipality with an alternative for convincing people in the area of project intervention to leave that area without becoming a budgetary burden to the Municipality.

34 Compliance criteria for OP 4.12 are: (i) participation of affected families throughout the resettlement process; (ii) compensation at minimum sufficient to restore their original living conditions; and (iii)

61

4.13 The Bank, as a condition for accepting the Piancó alternative, demanded: (i) consultations with affected families to gauge acceptability of the new site/conditions; (ii) a letter from the PMSL assuring the availability of 500 housing units earmarked for the Left Bank cohort (see PMSL letter of April 29, 2013); and, (iii) installation of individual hydrometers in the apartments, project-financed. However, in the interim, the PMSL had promised a major portion of available Piancó units to other families surveyed from other areas of the Municipality, stating that other buildings planned for Piancó would be used for project families. This significantly delayed the resettlement timetable and as a consequence, also delayed the important complementary urbanization/paving works planned for the Left Bank which also had environmental safeguards implications. Further, installation of hydrometers was far less than promised and some families were resettled in apartments without this equipment. Even after the Piancó option was agreed by the parties, the PMSL continued to delay moving forward in part due to the complex and heterogeneous profile of the targeted Left Bank families, a critically important factor not taken into account during Project preparation where most of the focus was on environmental/hydrologic issues and not on how the PMSL would operationalize the resettlement.35 4.14 Land use and control: The Project was intended to increase the municipality’s capacity to control and enforce land use to avoid new invasions and illegal sub-division of urban lands. The PMSL was slow to take action to control land use in the Project area including occupation and degradation of environmental preservation areas. Further, the Borrower’s unilateral changes to the Project concept and environmental approach (see Main Text 2.2) also affected resettlement and the Involuntary Resettlement Framework (IRF) prepared during preparation due to extensive hydraulic landfilling (aterros hidraulicos) to create new land for habitation around the Lake implying an increased number of families to be resettled, not covered under the IRF. Despite the Bank’s repeated insistence, the PMSL lacked the commitment and failed to adopt measures to demolish the homes of families relocated voluntarily, so that other families moved into those same houses creating a new, at-risk cohort. There was also evidence of a significant increase in new invasions of flood-prone areas and Permanent Protection Areas (APP), especially on the Left Bank, with no preventative or corrective actions. 36 Areas earmarked for implementation of the Das Bicas Urban Park Phase 2 were no longer viable due to uncontrolled occupation/invasion.

freedom of the affected population to choose between compensation alternatives. This Project met these criteria.

35 The major heterogeneity in the profiles of targeted Left Bank families - and in the quality and value of homes in the flood-risk area - had implications which were not taken into account at appraisal. Families living in a more precarious situation easily accepted the option of resettlement in an apartment building. But for a substantial contingent with more expensive/valuable homes and in less risky areas, the PMSL had to negotiate and persuade them to leave against their wishes (a major political burden) and was obliged to provide cash compensation (an economic burden). Another group of families in risk areas which did not adhere to the Piancó option were low-income families living in very poor housing whose way of life was unsuited to moving into an apartment building because they depended on animals for their livelihood and were thus not attracted to the PMSL’s offer. The PMSL could only convince, not impose, under voluntary adhesion principles.

36 Best practice resettlement normally favors “relocation by varredura” rather than “voluntary adhesion” as a means of formally releasing territory/land. In São Luís there is a large area on the Left Bank where flood-prone housing is located. Varredura is an approach through which people are withdrawn and resettled in a spatially sequential manner, giving the PMSL greater control over the use and occupation of land because in place of withdrawal from small, sparse areas, withdrawal can be handled in “sectors”. However, this approach can only be used for Involuntary Resettlement, so could not be used once the flood-risk cohort shifted to a voluntary basis. This issue was discussed repeatedly with the UGP but the

62

4.15 The overall resettlement oversight/enforcement burden exceeded UGP capacity. Efforts to establish a Social Unit in the UGP to monitor a set of activities including social activities envisaged in the works environmental manual, resettlement, and the sanitation and environmental education campaigns, encountered difficulties including political interference. Creation of this unit was late and it proved difficult to maintain the unit intact, i.e. with adequate professionals. 4.16 Indemnification and expropriation complexities also delayed resettlement of the works-affected families in the vicinity of the Rio das Bicas Canal on the Right Bank. The key fact distinguishing this resettlement as Involuntary is that the works supported by the Project required the re-taking of land by the State through its exercise of eminent domain. Delayed resettlement meant that the canal works could not be executed until families were moved out, resulting in a complex negotiation with the PMSL to secure a “social rent” to help families cover the cost of temporary accommodation until their new housing was ready, while permitting the canal works to proceed. Resettlement on the Right Bank followed all elements of the Bank’s policy including: identification of need; studies on urbanization and engineering designed to minimize the need for resettlement; and, negotiation with affected populations.

4.17 By 2014, primarily due to the PMSL’s weak execution capacity, constant delays and failure to comply with Bank recommendations, the social safeguards rating slipped to unsatisfactory. Up to end-2014, the failure to resolve the social situation of the flood-risk group and to reduce the number of families at risk constituted non-compliance with OP 4.12 because while the instruments were available, no action had been taken whether for political, capacity or other reasons. The Project however, did not take any actions which harmed families on the Left Bank and the lack of success over time to resolve the large-scale resettlement was one of poor implementation of Component 3, not non-compliance with safeguards since resettlement was converted to a voluntary basis. On the Right Bank the small number of families affected by the canal works were protected under the Involuntary Resettlement Plan/ Framework.

4.18 The ICR exit rating is Moderately Unsatisfactory with a risk rating of Substantial. Good progress was made in 2015 to implement the Resettlement Action Plans (RAP) for families affected by the Rio das Bicas canal works. Progress was more limited in the case of the large-scale resettlement under MCMV of families in Piancó, preventing about two thirds of eligible families from improving their living conditions within the life of the Project. However, while at closing it looked like the negative political and economic climate in Brazil would prevent successful completion of these activities post-closure, this assessment has changed. The Bank’s Safeguards Coordinator for Brazil and the Project Safeguards Specialist, under an agreement with the PMSL, will supervise completion of the involuntary resettlement. In the case of voluntary resettlement, the PMSL will periodically report progress to the Bank on physical completion and numbers of families resettled. Budget resources were allocated by the Global Practice. Further, an Action Plan including legal and budgetary guarantees for completing the Project’s resettlement-related activities was agreed with the PMSL and will be the basis for the Bank’s post-closure supervision. The status of resettlement at closing is summarized below:

latter decided to handle the Left Bank cohort through voluntary adhesion, which has distinct merits but does not permit the formal release of land for works, and complicates control over unoccupied land. As noted, the PMSL failed to demolish homes vacated by resettled Left Bank families, promoting further occupation of those same houses, and greatly delaying the release of those risk areas for project works, including environmental. The demolition issue was a source of tension between the Bank and PMSL. The Bank suggested and was willing to finance a contractor to do the demolition, but the PMSL insisted it had the capacity yet did not follow up.

63

Families affected by macro-drainage works: Project-supported macro-drainage works affected 52 families directly, who were resettled involuntarily in close compliance with the Involuntary Resettlement Plan. Of these, the assets of 22 families were partially affected and they received cash compensation (indemnification); the assets of another 27 families were totally affected and they were offered and received new housing units. Another three families were not affected by the works once design changes were made but due to their acute social vulnerability, the Client preserved an earlier, negotiated offer to relocate them to the new housing units. A total 33 new housing units were constructed for families who chose the new housing compensation alternative. At the ICR due date, 12 families had physically received their new homes, while another 21 families remained in temporary housing (receiving rent subsidies - aluguel social) until their units were ready (completion expected for May 2016). One landowner received cash compensation, the result of a court challenge.

Families resettled from flood risk areas: The Project financed improved housing for families living in sub-standard, flood-prone settlements along the Bacanga Lake and who voluntarily agreed to relocate. A new survey in 2012 found that 568 families (289 in flood-risk areas and 279 in other areas) would benefit from housing/urban improvements. The PMSL opted for a two-part strategy: reducing adverse impacts related to involuntary resettlement from these areas by inducing families to resettle voluntarily on the grounds that such a move would significantly improve their quality of life; and, preparing a Resettlement Action Plan for the remaining residents. When negotiations for Vila Embratel fell though, the PMSL accessed the Federal MCMV program, entering into an agreement with the Ministry of Cities (responsible for MCMV, which is 95% subsidized) to give priority to the Project’s target population. Several communication and mobilization campaigns promoted voluntary enrollment to which 525 families (108 and 417) responded. The first 108 were resettled, another 276 were screened and judged eligible and 141 await screening. About one-quarter of families in risk areas did not enroll due to concerns about livelihoods.

4.19 OP 4.37 Safety of Dams: While the Bacanga Dam is comparatively small, its importance to local land use and occupancy and to environmental conditions affecting large numbers of people, its use to connect the city to the Harbor and, visibly poor condition of the Dam resulted in a safety assessment study by an independent panel of experts during preparation (2007). Analysis of structural and operational conditions found strong evidence of deterioration and imminent collapse of key segments of the Dam. In February 2007, one floodgate collapsed. Due to the central importance of the Dam’s proper and efficient operation to the effectiveness of other project interventions, Dam rehabilitation was included in the Project including rehabilitation of its structure and electro-mechanical equipment which would eliminate the risk of rupture and permit adequate daily operation. Also contemplated were new operating rules to achieve a balance between controlling flooding and improving the quality of water in the Lake.

4.20 The Bank Environmental Safeguards Specialist’s analysis of the Dam experience and compliance provides an exit rating of highly unsatisfactory. Despite all efforts by the Bank and PMSL to resolve the impasse associated with the State’s management of the Bacanga Dam and legal obligation to rehabilitate it, the Project could not comply with OP 4.37, Safety of Dams, the rehabilitation works were not done and in September 2015, the main gate of the Dam collapsed. Despite the social and environmental hazards presented by the Dam in its deteriorated condition, and the body of technical studies, diagnoses, recommendations and detailed designs available to the State, it was/is not a State Government priority and the prospects for its rehabilitation appear slim. Compliance with this safeguard was consistently unsatisfactory in recent years and this rating was sustained at exit. The main factors preventing compliance with OP 4.37 are discussed below.

64

4.21 Rehabilitation of the Bacanga Dam was impeded by a protracted legal/political struggle between the State and Municipal Governments for overall operational responsibility. The Bank had sought since effectiveness to negotiate/facilitate its transfer to the PMSL and at the time of the MTR, prospects seemed better given that the State Infrastructure Secretariat (SINFRA) had issued an official document asserting that the PMSL should operate and maintain the Dam as a municipal facility. PMSL was authorized to finalize the transition through the Secretariat for National Patrimony (SPU), but no action resulted and the State retained control. 4.22 After years of inactivity during which the State failed to honor its commitments to the Project, 2014 saw some movement in the dam rehabilitation studies (but not coordinated with the PMSL), as well as the Dam Contingency Plan and the Operational Manual. In July 2014, the Bank team agreed with Bank Management to conduct an independent, expert legal and technical review of the Dam and a carefully-managed conversation with the State Government. However, despite the Bank’s repeated efforts to establish a closer dialogue with the State, especially the new administration, the Dam designs/procurement package for the works went out to bid without Bank review. The Bank’s insistence on seeing the designs resulted in only partial disclosure. 4.23 In May 2015, the Bank delivered the technical opinion of a Bank-contracted dam specialist and the project Environmental Specialist to the State Governor and the PMSL, rejecting the State’s technical proposal for rehabilitation and warning of the risks of pursuing the option of controlling tides through a single floodgate while repairs were executed. The State never responded to the Bank’s recommendation and did not conduct rehabilitation works on the Dam. In September 2015 the main floodgate collapsed. Emergency works to control the inflow of seawater – fortunately limited by the low tide and dry season – involved installing a large, temporary rock/earth barrier which was still in place at the time of ICR finalization. 4.24 The Project-financed Contingency Plan for the Bacanga Basin, prepared by the Municipal Secretariat for Security and Citizenship (SEMUSC) and outlining coordinated disaster-relief response actions by civil defense bodies, was not activated by the failure of the floodgate. The Bank team collected information on the impact of this Dam rupture and corrective measures taken, monitored the media, and kept Bank Management and the CMU informed, seeking guidance on potential actions to be taken. Clearly, political factors were the dominant influence throughout the seven-year experience but even with the new State/Municipal political alignment and signals of willingness to cooperate and move forward on the Dam, this was not the case. With the Project closing, and other crises dominating the State’s political discourse, the Dam was not a priority. 37

4.25 OP 4.04 Natural Habitats: Project interventions sought a significant positive impact on natural habitats through: adequate use and occupancy of the Bacanga region; maintenance of the maximum water level in the lake; carving out of a 30 meter-wide permanent preservation area/strip around the Bacanga Lake; and, a plan for the use of these areas (green areas/vegetation, urban parks, leisure areas) to promote their utilization by the population and prevent re-occupation. The Project also included proposals for a management instrument for the Bacanga State Park, a State Government responsibility.

4.26 The compliance rating was downgraded in late 2014 to unsatisfactory which continued through closing, based on a highly unsatisfactory rating for compliance with environmental

37 See: Parecer Tecnico sobre o Projeto Executivo de Restauração da Barragem Bacanga, 2015. In November 2015, after closing, SINFRA advised the Bank that due to lack of resources and emergency needs of the Dam, the State had abandoned the idea of total rehabilitation. Instead, the State would rehabilitate the stop-logs; strengthen the bridge structurally; and install one new floodgate. This plan was under discussion with the winning construction firm. Current status is unknown.

65

safeguards and moderately unsatisfactory for compliance with environmental laws. With the exception of the Management Plan for the Bacanga State Park – updated by the State (not the project) based on a 2009 version - and a partial and segmented permanent preservation strip with communal use in the area of the Canal das Bicas, none of the required interventions mentioned above was executed. No information is available on the extent to which land occupation issues or the envisaged recuperation of degraded areas in the interior of the above park, were resolved. According to the Borrower Completion Report, the greatest challenge facing this safeguard was in implementing natural habitat activities correlated with resettlement, and gains by the PMSL were partial. Each of the proposed activities is discussed briefly in the Annex 2 matrix.

4.27 OP 4.11 Physical Cultural Resources: The Project was not expected to affect areas of historical and cultural heritage in the City of São Luís as interventions were outside the Central Historic District. However, as the project entailed construction and excavation to expand and replace infrastructure, the Environmental Assessment (EA) included screening for known cultural property in the project area and incorporated chance finding procedures. In addition, Brazil has a well-developed legal and normative framework to deal with such issues, under the supervision of the National Institute for Historic and Artistic Patrimony (IPHAN). In practice, screening procedures for all works throughout the project period did not encounter any known cultural property and chance finding procedures were not invoked. This Safeguard was rated Satisfactory throughout.

66

Annex 5. Bank Lending and Implementation Support/Supervision Processes

(a) Task Team members

Names Title Unit Responsibility/

Specialty Lending

Megan Marie Bela Program Specialist (Temp) LCSUW-

HIS

Fernando Andres Blanco Cossio

Lead Economist GMFDR

Dean A. Cira Lead Urban Specialist GSURR

Martin P. Gambrill Lead Water and Sanitation Specialist (TTL, Preparation)

GWADR

Rogerio Dias Gerheim Consultant LCSDU

- HIS

Yasuhiko Matsuda Sr. Public Sector Specialist GSPDR

Claudio Mittelstaedt Cons. Financial Specialist LCSFM

- HIS

Alexandre Fortes Cons. Environmental Spec. LCSUW-

HIS

Liliana Pena Staff Assistant LCSOS -

HIS

Maria-Valeria Pena Lead Sociologist LCSSO -

HIS

Armando Pinheiro-Castelar Local Consultant CASPL

Paula Dias Pini Senior Urban Development Specialist

GSURR

Trajano Quinhôes Consultant LCSHH

- HIS

Julia Conter Ribeiro Senior Finance Assistant WFALN Wilson dos Santos Rocha Cons. Water Engineer GFADR Juliana Wenceslau Biriba Santos

Research Analyst LCSPS -

HIS

Luciano Wuerzius Sr. Procurement Specialist GGODR

Supervision/ICR

Ming Zhang Sr. Urban Development Spec (TTL)

GSU12

Oscar Alvarado Sr. Urban Development Spec (TTL)

GWADR

Emanuela Monteiro Urban Devt. Specialist (TTL) GSU10 Juliana Garrido Sr. Water/Sanit. Spec. (co-TTL) GWA04

Reynaldo E. Asturizaga Cons. Urban Specialist LCSUW

-HIS

Melissa Bonneton Junior Professional Associate LCSUW

-HIS

Joao Vicente Financial Management Specialist GGODR Peter Cohen Cons. Social Specialist GSU01 Catherine Lynch Senior Urban Specialist GSURR Wilson dos Santos Rocha Cons. Water Engineer GFADR

67

Eri Watanabe Cons. Social Specialist GPVDR Luciano Wuerzius Senior Procurement Specialist GGODR Alexandre Fortes Cons. Environ. Safeguards Spec. GWADR Luz Maria Gonzalez Cons. Economist GWADR Anna Roumani Cons. Operations Specialist GWADR Soraya Melgaço Cons. Social Safeguards Specialist GWADR Alberto Costa Senior Social Specialist GWADR Gilberto Canali Consultant GWADR

(b) Staff Time and Cost

Stage of Project Cycle Staff Time and Cost (Bank Budget Only)

No. of staff weeks USD Thousands (including travel and consultant costs)

Lending FY06 27.03 117.13 FY07 19.12 146.88 FY08 17.19 139.70 FY09 1.13 2.57

Total: 64.47 406.2838 Supervision/ICR FY09 13.90 75.25 FY10 19.16 125.39 FY11 13.53 116.99 FY12 18.66 85.13 FY13 24.36 83.55 FY14 21.71 91.75 FY15 24.77 94.59 FY16 9.94 38.50

Total: 146.03 711.15

38 The PAD shows total cost of preparation (at that time) of US$1.15 m, comprising Bank resources of US$350,000 and PHRD trust funds of US$800,000. The Bank’s updated system shows Bank resources totaling US$406,280, implying a final cost of US$1.21 m.

68

Annex 6. Beneficiary Survey Results The final beneficiary Satisfaction Survey was still awaiting procurement at the time of ICR conclusion. No results available.

69

Annex 7. Stakeholder Workshop Report and Results N/A

70

Annex 8. Summary of Borrower's ICR and/or Comments on Draft ICR

A. Summary of Borrower’s ICR: Informal Translation

Municipal Government of São Luís

Municipal Secretariat for Special Projects (SEMPE)

São Luís Enhancing Municipal Governance and Quality of Life Project

Program Management Unit

(Contract 7578-BR / Project P094315)

EXECUTIVE SUMMARY

The São Luís Enhancing Municipal Governance and Quality of Life Project was

formulated within the ambit of municipal public policies whose principal objective was

to improve the living conditions of the poor population inhabiting - in a totally

precarious manner – the region of the Bacanga River Basin.

Project activities were intended to disrupt the economic and social processes

sustaining urban expansion on the two banks of the Bacanga River, right and left, and

the advancing population on the low-lying mangrove areas, along the margins of the

Bacanga State Park and in flood-risk areas, thereby inverting the logic of occupation

and improving these lands.

The Project also considered that improving income would have an impact on

the physical configuration of the area and vice versa. From this point of view, indicators

such as the creation and improvement of parks and green areas (m²), creation and

requalification of public spaces (m²), requalification of the river front (m), creation of

cycling tracks (m), pedestrianization and physically improved transit on roadways (m²),

creation of new pedestrian routes (m), construction of footbridges and stairs (units), and

construction of social equipment (units), would gain importance.

The Program was structured in three main activity streams: (1) Strengthening

of Municipal Management and Local Economic Development (LED); (2) Basic

Sanitation and (3) Urbanization and Environment – through which were implemented

a set of multi-disciplinary and integrated activities designed to optimize benefits and

positive impacts of the Program for the population of more than 230,000 inhabitants,

in around 45,000 households in 45 regional neighborhoods.

71

The Bacanga Basin is the largest hydrologic basin totally contained within the

Municipality of São Luís. Besides the Bacanga River, the basin has as its principal

tributaries the das Bicas River, the Igarapé do Tapete, the Igarapé Itapicuraíba, Igarapé

do Tamancão and Igarapé do Piancó. The principal focus area of the program,

concentrated in the Bacanga Basin, was to include nine sub-basins: Vila Embratel,

Campus UFMA, Sá Viana, Pindorama, João Paulo, Filipinho, Coroadinho, Praia

Grande e Sacavém. In the Bacanga Basin, one finds important evidence of the city’s

history as well as cultural manifestations reflecting the rural-city relationship, such as

the Sítio do Físico ruins and the middens (sambaquis) found in the Bacanga State Park,

the Sítio Tamancão, today transformed into the Estaleiro School, the Festa da Juçara in

Maracanã, the staging of the Passion of Christ in the Anjo da Guarda neighborhood,

and the ruins of a Rice Mill/Factory, probably the oldest record of the history of local

industry.

The objective of the project was to improve the Borrower’s capacity in the

management, financial, urban, environmental and service-delivery areas with a view to

promoting local economic development and improving the quality of life of the

population living in the Bacanga River Basin.

To achieve this objective, the project had intermediate indicators sub-divided

among three components: Strengthening Municipal Management and Local Economic

Development (Component 1), Improved Water and Sanitation (Component 2) and

Urban and Environmental Improvements (Component 3).

Under Strengthening of Municipal Management and Local Economic

Development (Component 1) the objectives were: improve capacity and promote

local economic development, improve capacity and promote employment and income

generation in the project area, and improve municipal administrative capacity. These

objectives were achieved in the following way:

The preparation and execution of the Local Economic Development Plan (LED

Plan) brought about improved capacity and promotion of local economic development

through the preparation of short and medium-term planning for the Municipal

Government, resulting in the following products: (i) Multi-year Plan 2014/2017; and

(ii) Strategic Plan for São Luís 2033. Also prepared were: Portal for the Department of

Information and Economic Intelligence (www.diie.com.br); the Supplier Development

Program (SDP) whose integration with the State Government system is under

discussion, and will promote the identification of suppliers in an easier form

72

(prioritizing the health, education and food areas); and, notably, the training received

by the Municipal Government team.

Also developed was a Tourism Cluster Plan which financed training activities,

preparation of a marketing plan and a São Luís city “brand”, and benchmarking trips

for industry personnel. Also executed were: a tourist map; a diagnosis of human

resource qualifications in the tourism industry; and, on a pilot scale, training of tourism

managers.

To improve capacity and promote the generation of employment and income in

the project area, an Innovation Plan was developed which trained individual

professionals and micro-entrepreneurs, reaching a total of 540 persons. There was also

an agreement with SINE and with the Civil Construction Union, to train 676

professionals in the field of civil construction.

To improve municipal administrative capacity consultancies were conducted

which resulted in the Strengthening of the Municipal Procurement System including

assistance to the Central Procurement Committee and a course to disseminate fiscal

information. Further, the Program Planning and Management System was

implemented (PPMS). The primary function of this system is to provide a tool to the

Program Management Units in public entities enabling their efficient and effective

management of the activities promoted by them.

To achieve Improved Water and Sanitation (Component 2), the objectives

were improvements to sewerage systems, urban drainage systems and water supply

systems.

These objectives were partially achieved through improvements in urban

drainage systems with the implementation of works for two macro-drainage canals: the

Coroado Canal with an extension of 368 meters benefiting 7,241 residents; and the

Córrego Salinas do Sacavém and Rio das Bicas Canal with a length of 2,095 meters,

benefiting 12,272 residents.

In the case of improvement of the sewerage system and water supply system,

these objectives were not achieved, as the objective was to implement the sewerage

system on the Left Bank of the Bacanga Basin which would feed the Bacanga

Treatment Plant and expand the network on the Left Bank of the Bacanga Basin to the

pressure zones of VI-A - Vila Embratel and VI-B – Anjo da Guarda. Both were

counterpart-funded investments using resources of the Federal Government’s Growth

Acceleration Program (PAC), which were not executed in their totality and thus did not

73

generate significant results. These works were not executed within the period of the

Loan Agreement due to failures in their Final Designs (which were already revised)

and due to Brazil’s economic situation which created delays in the transfer of approved

PAC resources.

In regard to the Urban and Environmental Improvements (Component 3),

the objectives to be achieved were the resettlement of families threatened with flooding

related to tidal movements on the Left Bank of the Bacanga Basin, rehabilitation of

urban roads in the area of the project, complete rehabilitation and effective operation

of the floodgates of the Bacanga Dam, development and approval of a Municipal Basic

Integrated Sanitation Plan, regularization/registration of properties held by the local

population, and creation of public spaces (green areas, urban parks etc).

The Municipal Government resettled 108 families on the Left Bank of the

Bacanga Basin in a housing complex (Piancó VII and VIII) financed by the Federal

Government’s Minha Casa Minha Vida Program, and conducted a survey/census of

another 487 families also on the Left Bank for resettlement in Piancó I and II. At the

completion of the ICR, Banco do Brasil had approved the documentation for 263

families. Their information is being updated in the bank’s database, for further contract

signing. The estimated date for the delivery of the housing units is May 2016.

Also executed were 5.42 Km of paved roads with Federal Government PAC

resources and 1.1 Km with World Bank loan resources. Public spaces totaling

14,932.09 m² were also created.

The Municipal Integrated Basic Sanitation Plan was prepared and is currently

before the Municipal Attorney General for conveying the Law to be approved by the

Municipal Chamber of Councillors.

The Municipal Government of São Luís is carrying out, in partnership with the

Federal University of Maranhão, land regularization through a grant of property title to

families of the Sá Viana neighborhood on the Left Bank of the Bacanga Basin.

The State Government of Maranhão prepared an executve project (final design)

and put out for bidding works for the Rehabilitation of the Structure of the Bacanga

Dam. The Service Order to start the works has not yet been signed.

By the end of the Loan Agreement, the Municipality of São Luís achieved

impressive results from the implementation of this project. On the Right Bank, one can

observe macro-drainage works of the Coroado and Rio das Bicas Canals which are

limiting the flooding occurring in the neighborhoods of Pindorama, João Paulo,

74

Filipinho, Coroadinho and Sacavém. Besides this, there has been an improvement in

the socio-environmental quality of the region in the vicinity of these works with

urbanization through the Rio das Bicas Urban Park and paving of roads around the

canals, as well as a result of the resettlement activities. The Resettlement Plan for the

Right Bank, whose initial objective was to compensate the families with homes totally

affected by the Project works, also ended up improving the quality of life not only of

these families, but of the overall families living in the nearby vicinity. The improved

quality of life of the communities is acknowledged, as well as their increased perception

of security and their development of small businesses.

Also on the Left Bank, while not showing such evident results, urbanization

works comprising two plazas and a Football Stadium pertaining to the Requalification

of the Left Bank works, were important activities which have benefited the population

resident in the area.

The big conquest on the Left Bank was the voluntary resettlement of families in

risk areas. Using a different approach than the one used for the Right Bank, the

resettlement activities in the Left Bank took into account 597 families. In a first stage,

the Program managed to resettle 108 of these in housing units which are part of a Minha

Casa Minha Vida financed complex located near the original houses. Through a long

process in partnership with the community, the Project/UGP reached out to families,

asking them whether they wanted to voluntarily adhere to resettlement. Under a second

stage, 487 families applied for voluntary resettlement. Among the cadasters sent to

Banco do Brasil for approval, 263 were cleared for further contract signing and receipt

of the key to a new house. There is a remaining deficit against the original resettlement

target in the Left Bank. However, through Project implementation, it became clear that

new forms of working out these resettlement issues are possible (taking into

consideration the variety of existing situations) and reaching a larger number of families

at risk.

To develop and implement this project, about 60% of the value of the Loan was

utilized and 6.3% of counterpart resources were executed.

Positive and Negative Factors affecting the Project:

Despite the low financial disbursement, some positive and negative factors

contributed to the moderately satisfactory performance of the project.

Positive factors included: (1) the creation and training of a Special Procurement

Committee (SPC) to conduct procurement in accordance with World Bank rules and

75

training of this Committee; (2) training and technical visits of the Project Management

Unit (PMU) to other World Bank projects during the preparation and initiation phase

of the Project provided the opportunity to see close up the methodology for

implementing programs; (3) support from the community in the preparation of the final

designs, as well as the availability of World Bank consultants to help the PMU to

prepare projects and take decisions saved time in finding responses to technical

questions which arose during implementation; and, (4) the increased frequency of

World Bank missions and technical visits by Bank consultants in the final phase of

implementation also helped, along with agile payment processes for World Bank

resources.

Negative factors included: (1) the complexity involved in project design which

included two geographically and socially distinct regions and the dispersion of multiple

activities, without a specific focus in one area or type of intervention; (2) changes in

the coordination of the project throughout project preparation and implementation,

program activities which involved other public entities and institutional capacity in

relationships/communication with other organs of the Municipal Government also

played a role; (3) poor quality of the counterpart-financed projects/works and the time

lapse between preparation of projects and their initiation, along with delayed transfer

of Federal resources, were other negative factors; (4) the lack of a consultant to assist

the PMU in managing the Program was also a negative factor.

Besides the above, on various occasions, problems arose such as the delayed

responses of the World Bank to requests for No Objection and divergent

opinions/understanding within the Bank team in regard to No Objections.

Another factor which contributed to project performance was the restructuring

of the Project in 2013 which included activities which ended up not materializing within

the closing date of the Loan Agreement, even though targets which could not be

achieved were withdrawn from the program, and targets which were believed to be

completely feasible were included. Also, even though the program already had Federal

resources guaranteed and projects ready, these targets were not achieved. This fact

was due principally to the worsening of the national economy - with the delayed transfer

of Federal resources - and to inconsistencies in works which created delays and

paralysis. Therefore, the restructuring contemplated was not sufficient to improve

project performance.

76

The PMU gave priority to Loan-financed works which substantially improved

the quality of life of residents of the region, bringing improvements to the urban space,

environmental improvements, increased security, development of small businesses etc.

Lessons Learned:

By the end of project implementation, achievements and errors can be identified

which provide lessons to be learned by the Municipality in order to contract a new Bank

Loan, such as: (1) delimit an area of project coverage more concentrated geographically

and with more co-related activities; (2) do not have targets for activities which depend

directly on other federal or state entities and instead, have activities which depend only

on the entity taking the Loan; (3) start the program with final designs ready and/or

execute projects/works necessary for structural interventions at the beginning of the

project during preparation or in the first year, and the three or four final years of

implementation should be dedicated to the execution of works; (4) a project manager

should be contracted at the beginning of the program so that there will not be a loss of

information over the course of project implementation even if there are changes in the

PMU team; (5) also, one should propose areas of intervention free of problems of

expropriation, because resettlement problems create major delays in the execution of

some works and prevent the execution of others; and, in conclusion, (6) there should be

included in the project and/or in contracts, if possible, financial resources for

indemnification, since difficulties exist for the Municipal Government to take

responsibility for financing such action.

77

B. Borrower Comments on Draft ICR:

80

Annex 9. Comments of Co-financiers and Other Partners/Stakeholders N/A

81

Annex 10. List of Supporting Documents Project Concept Note (PCN) and Project Information Document (PID) Project Appraisal Document (PAD) Integrated Safeguards Data Sheet (ISDS) Avaliação Ambiental Regional – Município de São Luís (Resumo Executivo), MMT Planejamento, June 2007 Loan Agreement Operational Manuals (2007, 2011) Restructuring Paper (2013) Supervision Aide Memoires Technical Papers (environmental, social, urban) Implementation Supervision Reports (ISR) Project Progress Reports Procurement PPRs and Aide Memoires Financial Management Reports and Aide Memoires Audit Reports Country Partnership Strategy (2008-2011, 2012-2015) Brazil Systematic Country Diagnosis: Realizing Brazil’s Potential and Fulfilling its Promises, World Bank, Report #101431, 2016. Baseline Satisfaction Survey (2012) Nota Técnica: Salvaguardas Ambientais, A. Fortes, September 2011. Relatório de Conclusão do Programa Bacia do Bacanga, SEMPE/UGP, March 2016 Local Economic Development Strategy, SEBRAE, 2012 Plano de Inovação da Bacia do Bacanga, SEBRAE, June 2012 Plano de Ação: Plano de Inovação da Bacia do Bacanga, SEBRAE, 2012 Mercado de Trabalho para Mulheres Habitantes da Bacia do Bacanga (Município de São Luís, MA). V. A. Saboia, August 2010

82

Gênero, Trabalho e Violência: Inserção de Mulheres no Mundo do Trabalho, V. A. Saboia, February 2011 Plano de Negócios para Capacitação e Intermediação de Mão de Obra Feminina oriunda do Território da Bacia do Bacanga para atuar nas Obras do Programa, F. de Holanda, January 2011 Plano de Desenvolvimento do Cluster de Turismo de Sào Luís – Versão Final, Sociedade Portuguesa de Inovação (SPI), July 2011 Projetos de Desenvolvimento do Cluster de Turismo de São Luís – Versão Final, SPI, June 2012

83

Annex 11. Photographic Record of Project Investments

1. CANAL DO COROADO (Right Bank of the Bacanga River)

BEFORE:

The areas around the canal were constantly subject to flooding.

AFTER:

The Project supported macro drainage works. There are ongoing counterpart investments in the

implementation of a multipurpose cultural and sports center on the same site (Praça CEU).

84

2. CANAL RIO DAS BICAS (Right Bank of the Bacanga River)

BEFORE:

The areas around the canal were constantly subject to flooding.

AFTER:

The Project supported macro drainage works, as well as the urban upgrading and

implementation of amenities in the canal area (see item 3 below, on Parque das Bicas).

85

BEFORE / AFTER:

People affected by the canal works on the Right Bank of the Bacanga were resettled in

new homes built in the close vicinity.

86

3. PARQUE DAS BICAS (Right Bank of the Bacanga River)

BEFORE:

The area was constantly flooded and lacked urban amenities.

AFTER:

In addition to the macro drainage works, amenities were implemented along the canal and

in various pockets where there was open space.

87

The urban upgrading works included the implementation of squares and sports courts,

street paving, lighting and sidewalks along the canal stretch.

88

4. VOLUNTARY RESETTLEMENT OF FAMILIES LIVING IN RISK PRONE AREAS (Left Bank of the Bacanga River)

BEFORE:

Areas in the Sá Viana neighborhood were constantly subject to flooding, in addition to

being at risk due to the poor operation of the Bacanga Dam.

AFTER:

Areas in the Sá Viana neighborhood were constantly subject to flooding, in addition to

being at risk due to the poor operation of the Bacanga Dam. The flood-risk families were resettled in apartment buildings at Piancó, funded by counterpart investments (Minha Casa Minha Vida).