Good Practice Framework - World Bank

PROBLEM-DRIVEN

GOVERNANCE

AND POLITICAL

ECONOMY ANALYSIS

Good Practice Framework

PROBLEM-DRIVEN

GOVERNANCE

AND POLITICAL

ECONOMY ANALYSIS

Good Practice Framework

Verena FritzKai KaiserBrian Levy

September 2009

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

WB456286

Typewritten Text

82652

PROBLEM-DRIVENGOVERNANCE AND POLITICALECONOMY ANALYSIS

GOOD PRACTICE FRAMEWORK

SEPTEMBER 2009

©2009 The International Bank for Reconstruction and Development / The World Bank1818 H Street NWWashington DC 20433Telephone: 202-473-1000Internet: www.worldbank.orgE-mail: [email protected]

All rights reserved

This volume reflects the view of the authors. The findings, interpretations, and conclusions expressedin this volume do not necessarily reflect the views of the Executive Directors of The World Bank orthe governments they represent.

The World Bank does not guarantee the accuracy of the data included in this work. The bound-aries, colors, denominations, and other information shown on any map in this work do not implyany judgment on the part of The World Bank concerning the legal status of any territory or the en-dorsement or acceptance of such boundaries.

Rights and Permissions

The material in this publication is copyrighted. Copying and/or transmitting portions or all of thiswork without permission may be a violation of applicable law. The International Bank for Recon-struction and Development/The World Bank encourages dissemination of its work and will nor-mally grant permission to reproduce portions of the work promptly.

For permission to photocopy or reprint any part of this work, please send a request with com-plete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com.

All other queries on rights and licenses, including subsidiary rights, should be addressed to theOffice of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: [email protected].

Cover photos, left to right: Picture 1 (protest): Panos Pictures Picture 2 (electricity): Jacob Silberberg/Panos Pictures Picture 3 (voting): Sanjit Das/Panos Pictures

Contents

iiiP R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

EXECUTIVE SUMMARY VII

ABBREVIATIONS XV

INTRODUCTION: WHY GOVERNANCE AND POLITICAL ECONOMY ANALYSIS? 1

PART ONE: A PROBLEM-DRIVEN APPROACH TO GOVERNANCE AND POLITICAL ECONOMY 5I.1. Three steps of problem-driven GPE analysis 7I.2. Good enough governance and feasible approaches to reform 10

PART TWO: TRANSLATING ANALYSIS INTO ACTION: USING PROBLEM-DRIVEN GPE ANALYSIS TO

ENHANCE BANK STRATEGIES AND OPERATIONS 15II.1. Informing Bank teams and adjusting strategies and operations to an existing

space for change 17II.2. Enhancing and broadening policy dialogue and developing innovative

approaches to operations 18II.3. Using GPE analysis to develop campaigns for change proactively 19

PART THREE: CHOOSING THE LEVEL OF ANALYSIS: COUNTRY, SECTOR, PROJECT 23

PART FOUR: EVIDENCING PROBLEM-DRIVEN GPE ANALYSIS 27

PART FIVE: GETTING THE PROCESS RIGHT: ISSUES TO CONSIDER 31V.1. Planning GPE work and linking it to operations or technical analysis 31V.2. How to define and find the necessary skills 32V.3. Implementing GPE diagnostic work and ensuring quality management 34V.4. Sharing/disseminating outputs 36V.5. Moving from analysis to follow-up 37

PART SIX: CONCLUSION AND LOOKING FORWARD 39

ANNEX 1: POLITICAL ECONOMY: KEY CONCEPTS AND APPROACHES 41I.1. Three clusters of drivers: structures, institutions, actors 41I.2. Capturing historical legacies 44

iv P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

C O N T E N T S

I.3. Institutional and stakeholder mapping and analyzing informal institutions 45I.4. Resources, rent distribution, political stability, and legitimacy 47I.5. Patronage networks, clientelism, and neo-patrimonialism 50I.6. Selected key concepts from institutional and political game theory 51

ANNEX 2: PGPE ANALYSIS AT THE COUNTRY, SECTOR/THEMATIC, AND PROJECT LEVELS 53II.1. ‘Politics in action’ at the country level 53II.2. Sector-focused and thematic problem-driven GPE analysis 60II.3. Problem-driven GPE for specific projects or single policy decisions 69II.4. Summary: Levels of analysis 70

REFERENCES 71

BOXES

Box 1: Using governance indicators and governance surveys as inputs to GPE analysis 6Box 2: Defining feasible solutions—Zambia and Mongolia 9Box 3: GAC analysis shaping the CAS—Zambia and Uzbekistan 17Box 4: Mongolia—developing an innovative approach to Mining Sector TA 19Box 5: The Philippines—building a coalition for procurement reform 20Box 6: Dissemination experiences of pilot teams 36Box 7: Moldova Governance Assessment—migration and in-country stakeholder and

institutional dynamics 43Box 8: Formal and informal institutions 45Box 9: Analyzing winners and losers of electricity sector reforms in India 46Box 10: Expected utility stakeholder analysis 48Box 11: South Sudan GAC for CAS background note 54Box 12: Country-level analytical questions developed for the GAC for CAS pilots7 55Box 13: Electoral incentives in Benin and the lack of growth-focused policies 58Box 14: Analyzing policy-making processes in Latin America 59Box 15: Governance and political economy analysis informing public sector reform

operations—Moldova and Afghanistan 63Box 16: Sample questions for sector-focused GPE analysis 65Box 17: Examples of PGPE for service delivery sectors—transport in Bangladesh, electricity

in Lebanon 66

vP R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

C O N T E N T S

FIGURES

Figure i: Three layers of problem-driven GPE analysis ixFigure 1: Three layers of problem-driven GPE analysis 7Figure 2: Options for translating GPE analysis into action 16Figure 3: Levels of analysis: country, sector, project 24Figure 4: Three clusters of drivers 42Figure 5: Macro and micro political economy interactions in core governance reforms 62Figure 6: Using value chains as a tool to disaggregate sector components 64Figure 7: Sector value chain for Natural Resource Management 68Matrix 1: Levels of analysis and key skills needed 33Matrix 2: Types of consultants for problem-driven GPE analysis 35

P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I Svi

Acknowledgments

The authors would like to thank Sanjay Pradhan(Vice President, WBI), Randi Ryterman (Director,WBI) and Debbie Wetzel (Director, PREM), whoprovided overall guidance for this work. Commentsand helpful suggestions were received from RosaAlonso Terme (EASPR), Shahrooz Badkoubei(PRMPS), Naaz Barma (EASPR), Sabine Beddies(MNSSO), Genevieve Boyreau (EASPR), RuxandraBurdescu (PRMPS), Alessane Diawara (AFMML),Chris Finch (EACCQ), Zahid Hasnain (EASPR),Motoky Hayakawa (AFTPR), Joel Hellman(SASGP), Stefan Hochhuth (PRMPS), Kapil Kapoor(AFMZM), Phil Keefer (DECRG), Asmeen Khan(EAPVP), Stuti Khemani (DECRG), Maks Kobon-baev (PRMPS), Jana Kunicova (ECSP4), TonyLambino (EXTOC), Sumir Lal (EXTOC), NickManning (LCSPS), Andrew Norton (SDV), Bar-bara Nunberg (EASPR), Sina Odugbemi (EX-TOC), Anand Rajaram (AFTPR), Francesca Reca-natini (PRMPS), Cathy Revels (ETW), ArshadSayed (EACMF), Joel Turkewitz (EAPVP), CabyVerzosa (WBIGV), and Alison Wescott (WBIGV),as well as Stefan Kossoff (DfID), Renate Kirsch(GTZ), and Doris Voorbraak (Dutch Ministry ofForeign Affairs/Development Cooperation).

The authors are especially grateful to the threepeer reviewers Ed Campos (Adviser, WBI),Stephen Ndegwa (Lead Specialist, AFTPR), and

Yasuhiko Matsuda (Senior Public Sector Specialist,EASPR). All remaining errors are solely the re-sponsibility of the authors.

Many thanks go to Madjiguene Seck (PRMVP)and Alice Poole (PRMPS) who supported the fi-nalization and publication of this report, as well asto Humaira Qureshi (PRMPS) for her excellentadministrative assistance.

We gratefully acknowledge the support of theBank-Netherlands Partnership Program (BNPP)under a PREM-Governance window award for‘Developing Strategies for Addressing GovernanceConcerns In Challenging Country Contexts(Breakthrough Facility)’ without which work forsupporting initial demonstration cases and devel-oping the framework would not have been possible.

Authors’ contacts:

Verena Fritz (Governance Specialist, PRMPS)

[email protected]

Kai Kaiser (Senior Economist, PRMPS)

[email protected]

Brian Levy (Adviser, PRMPS)

[email protected]

Executive Summary

P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

Governance and political economy analysis has acrucial part to play in enhancing the effectivenessof development. Across the World Bank, there isa widely shared perception that we need to gaina better understanding of the environments inwhich Bank operations are taking place andseek to promote progressive change (IEG 2006,2008). This coincides with an increasing recog-nition that governance and political economy(GPE) factors play a powerful role not only in acountry’s overall development path, but also forshaping policies in various sectors and the waythey are being implemented. Moreover, an an-alytic approach to governance and politicaleconomy is essential to making progress in ad-dressing governance to unlock development, ascommitted to in the 2007 GAC Strategy andsubsequent Implementation Plan.

The objective of this good practice frame-work is to systematize approaches to governanceand political economy analysis and to providereadily available orientation for World Banktask team leaders (TTLs) and teams. This frame-work does not set out a particular product. Ittakes the general view that in order to improvedevelopment effectiveness, GPE diagnosticsshould become integral to preparing and im-

plementing Bank strategies and operations. Thekey intention of the framework is to proposestandards for such diagnostics and to synthesizelessons learned.

The framework is primarily intended for aWorld Bank audience. It is important for WorldBank management and staff to devote greater at-tention to the governance and political economydimensions of development challenges. This canbest be facilitated by a discussion that is tailoredto the needs of such an audience. However, thisis also a public document, and many of theideas and issues set out here may be of interestto stakeholders in client countries (such as re-searchers and civil society groups), and to otherdevelopment agencies and multilateral devel-opment banks (MDBs).

This good practice framework is one key pil-lar of an internal Sharepoint web resource(http://connect.worldbank.org/units/prem/PD-GPEA/PGPE/default.aspx). The web site makesthe range of approaches and applications whichthis framework synthesizes available in full. Italso provides a platform for sharing GPE ana-lytic products with other Bank staff. The frame-work and the site are intended to facilitate accessto a variety of approaches to governance and

vii

political economy analysis, as well as to facilitateexchange and feedback around existing andplanned work. It is intended to support anemerging community of practice (CoP) for GPEanalysis that spans different units and regionsacross the Bank.

WHY GOVERNANCE AND POLITICALECONOMY ANALYSIS?

Politics and political economy influencewhether and how reforms happen—in devel-oping as well as developed countries. For manyyears, World Bank teams have experimentedwith various ways of analyzing and understand-ing the political economy context of reforms andwith using such diagnostics for smarter engage-ment.1 The good practice framework describedhere is an attempt to summarize relevant analytictools and approaches and to make key lessonslearned readily available. It also seeks to set outhow such tools can be used in a way that is prob-lem driven, that is, focused on specific issues andchallenges rather than on developing broadoverviews, in order to generate useful findingsand implications.

GPE analysis will not lead to quick fixes,but it can be practical and useful for enhanc-ing strategies and operations. GPE analysisshould not be expected to deliver a new magicbullet. At the same time, recent experienceshows that it can be practically oriented andvaluable, setting out options and solutions ratherthan mainly pointing to obstacles. Across theBank, country and task teams can benefit sig-nificantly from a more systematic understandingof the context they are operating in and in whichthey are seeking to foster change.

The intention of this framework is to presenta menu of options rather than to offer or pre-scribe one particular way of doing things. Itseeks to provide guidance that is practical andspecific but which can and should be tailored toparticular circumstances and specific questionsto be addressed.

The emphasis of this paper is on good diag-nostics. This is complemented by a brief dis-cussion in Part Two concerning how analysis canbe translated into action. This aspect will beexplored more fully in an accompanying paper.A further paper is expected to address the im-portant issue of how country teams can monitorthe evolution of political trends more systemat-ically on a continuous basis in order to be able torespond to continuously evolving situations.

A PROBLEM-DRIVEN APPROACH TOGOVERNANCE AND POLITICALECONOMY ANALYSIS

The emphasis of this framework is on ‘prob-lem-driven’ analysis. It emphasizes GPE analy-sis that focuses on particular challenges or op-portunities, such as analyzing why reforms in thepower or health sector or those aimed at im-proving urban development might not havegained traction and what could be done differ-ently to move forward. ‘Problem-driven’ doesnot mean focusing exclusively on areas of diffi-culty. For example, in a range of environmentsmuch can be learned from including an analy-sis of how islands of excellence emerged. Inother cases, the key issue may be how to react toa ‘window of opportunity’ that is opening up.

The use of various types of GPE analysis toassess risks and to help shape reforms is becom-

E X E C U T I V E S U M M A RY

viii P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

1. While the World Bank’s mandate explicitly precludes it from engaging in politics, it has become widely accepted that the in-stitution needs to understand the political economy context of reforms from a diagnostic perspective in order to be able to as-sist countries effectively in designing and implementing development strategies and policies.

ing increasingly widespread across the privateand public sectors.2 The OECD is adopting aprogram on ‘Making Reforms Happen’, in-tended to support member countries in betteranalyzing the political economy factors in re-forms and to assist them in building successfuladvocacy and coalitions for reform.3 Manydonor agencies are seeking a better understand-ing of GPE contexts. DfID, the Netherlands,Sweden, and the EU are especially investing inthis area. Around the Bank, a number of Insti-tutional and Governance Reviews (IGRs) andPoverty and Social Impact Assessments (PSIAs),as well as analytical work carried out as part ofcountry, sector, and project GAC processes, haveincluded substantial political economy content.This framework builds on these existing ap-proaches and pilots.

Three layers of problem-driven analysis

A problem-driven approach to GPE analysiscomprises working through three layers: (i)identifying the problem, opportunity or vul-nerability to be addressed, (ii) mapping out theinstitutional and governance arrangements andweaknesses, and (iii) drilling down to the polit-ical economy drivers, both to identify obstaclesto progressive change and to understand wherea ‘drive’ for positive change could emerge from.This basic approach can be applied to analysis atcountry, sector, or project levels.

As outlined in Figure i, the first layer requiresdefining the challenge to be addressed (and to es-tablish that it appears to have a governance orpolitical economy dimension).4 Often, this will


P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S ix

F i g u re i : Three layers of problem-driven GPE analysis

Whatvulnerabilities/

challenges?

Evidence of pooroutcomes towhich GPEweaknessesappear tocontribute

E.g.: repeated failure to adopt sector reforms.Poor sector outcomes. Infrastructure identifiedas constraint to growth but not effectivelyaddressed. Continuous food insecurity.Corruption continues to undermine the businessclimate even after anti-corruption law

GPE

an

alys

is

Institutional &governancearrangements& capacities

What are theassociatedinstitutional set-up andgovernance arrangements?

Mapping of relevant branches of goverment,ministries, agencies, and SOEs and theirinteraction. Existing laws and regulations. Policyprocesses (formal rules and de facto). Whatmechanisms intended to ensure integrity andaccountability and to limit corruption exist?

Politicaleconomydrivers

Why are thingsthis way? Whyare policies orinstitutionsarrangements notbeing improved?

Analysis of stakeholders, incentives, rents/rent distribution, historical legacies, and priorexperiences with reforms, social trends andforces (e.g., ethnic tensions), and how theyshape current stakeholder positions and actions

Pro

ble

m-d

riv

en

2. For example, see http://www.pwc.com/extweb/pwcpublications.nsf/docid/EAB01AC994713716852570FF006868B6.3. See: http://www.oecd.org/document/15/0,3343, en_ 2649_34487_41399503_1_1_1_1,00.html.4. In many cases, this is not about identifying governance as the main overall bottleneck to development. Rather, it means con-

sidering whether other development challenges—such as improving infrastructure or getting traction on pension systems re-form or managing urban development—have governance and political economy dimensions to them that need to be addressed.

emerge from ongoing policy dialogue or existingreports. The second layer aims at understandinginstitutional and governance arrangements andhow these are related to poor outcomes. Thethird layer aims at drilling down to the under-lying political economy drivers.

Layers two and three clearly overlap. How-ever, they are differentiated in order to empha-size that institutional and governance dimen-sions as well as stakeholders and their interestsneed to be explicitly considered. Analysis atlayer two is essential for identifying what reformsare feasible from an institutional perspective.Drilling down to the political-economy layer isimportant to understand why the identifiedproblem has not been addressed successfullyand what the relative likelihood is of stakeholdersupport for various change options.

‘Good enough’ governance andfeasible approaches to reform

The approach to GPE analysis proposed here isbased on a diagnostic rather than a prescriptiveapproach to governance (Rodrik 2008a and2008b). The ultimate goal is to achieve devel-opment and improved governance, but the fo-cus is on how to move in this direction ratherthan on the final goal. ‘Good enough gover-nance’ captures the notion of focusing on pri-orities and improvements that are feasible (andaffordable) rather than trying to reform gover-nance wholesale—whether at the macro, core, orsector level (Grindle 2007, Khan 2006).

In many cases, the implications resultingfrom a GPE analysis will entail considering fea-sible options for reforms. While ‘first best solu-tions’ are technically superior in a textbooksense, their advantage can be lost or even re-versed under real-world conditions if they areonly partially implemented, are stalled or re-versed because they are politically unpalatable,

or trigger unintended consequences during im-plementation. At the same time, ‘feasible’ reformoptions are clearly distinct from ‘default op-tions’, that is, the reform path or non-reformpath that would result from no intervention oreffort at change.

TRANSLATING ANALYSIS INTOACTION: USING PROBLEM-DRIVENGPE ANALYSIS TO ENHANCE BANKSTRATEGIES AND OPERATIONS

Problem-driven GPE analysis can be trans-lated into action in several ways. This prob-lem-driven approach to analysis (hereafter re-ferred to as PGPE analysis) can provide politicalintelligence for agreement on what is feasiblewithin teams and with management. Moreover,it can provide advice on shaping strategies andoperations in ways that range from adjustingthem to the existing space for change to devel-oping pro-active strategies for expanding thespace for change. Country situations matter forcalibrating what action to take.

A basic way of using GPE analysis is to haveit inform country strategies and operations. GPEanalysis helps to create a better understanding ofthe environment in which operations take placeand in which strategic results are being pur-sued, and helps spell out crucial governance andpolitical (economy) risks. If context permits,GPE analysis can contribute to enhancing thepolicy dialogue and to developing innovativeapproaches to operations.

PGPE analysis can support innovative ap-proaches to operations. For example, teamsmay recognize that it is important as well as fea-sible to promote a better informed public debateon reform issues and policy options. Or it maybecome apparent that monitoring efforts by lo-cal stakeholders can be encouraged and sup-ported in order to limit corruption and achieve


x P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

results more effectively. To pursue such innova-tive approaches, it is vital that technical andGPE-focused diagnostic work be well integratedand that management attention be supportive.

PGPE analysis can also serve as a basis fordeveloping support for proactive campaignsfor change, for example, for getting seriousabout on improving health services or improv-ing infrastructure construction and mainte-nance. The main avenues for building tractionfor change are coalition building, informationand communication campaigns, and generallyintensified work with stakeholders on both thesupply and the demand side. PGPE analysis, assuch, can help decide whether it is possible topursue a proactive strategy of building momen-tum for reforms. Proactive strategies need to bedeveloped with respect for the preferences andchoices of local stakeholders and with the Banktaking the role of an ‘honest broker’. Impor-tantly, such approaches require resources, dedi-cated staff, and management attention.

POTENTIAL LEVELS OF ANALYSIS:COUNTRY, SECTOR, PROJECT

PGPE analysis can broadly be applied at threelevels: the macro or country level, the sector andthematic level, and the project and policy-spe-cific level. PGPE analysis can focus at one levelonly or cover several levels, usually through a setof papers.5 Teams engaged in strategy planning,for example preparing a country assistance strat-egy (CAS) or a major development policy loan(DPL) or re-engaging with a country, may seek tocover the country level as well as one or several keysectors. Task teams working on one particularoperation will be most interested in sector- and/orproject-focused GPE analysis—which ideallycan draw on existing country-level analysis.

Country-level analysis seeks to capture theoverall governance situation and the main po-litical economy drivers. Analysis at this levelserves to capture important factors such as thegeopolitical context, important social divisions(including ethnic or religious ones where rele-vant) and how they have evolved over time, andthe evolution of the political management ofeconomic rents. Sector and thematic PGPEanalysis seeks to analyze institutional and gov-ernance arrangements and political drivers in aparticular sector or subsector (covering an entiresector ‘value chain’ [see Annex 2] or selected is-sues within a sector or theme). Problem-drivenGPE analysis for specific projects and/or policydecisions addresses specific questions regardingproject design and management, or may be fo-cused around the political economy and insti-tutional aspects of an intended policy changeand its likely impacts (including GPE analysis aspart of PSIAs).

EVIDENCING PROBLEM-DRIVEN GPE ANALYSIS

Credible GPE analysis needs to be well evi-denced. It needs to pull together a compelling‘analytic narrative’ consistent with experienceand systematically gathered data and informa-tion—while avoiding the pitfall of essentiallyproviding technical analysis combined withsome broad statements about governance andpolitical economy. In particular, for sector andproject-focused diagnostics, PGPE work shouldbe closely integrated with technically focusedanalysis and teams, so that the two can be com-plementary and so that important synergies canbe achieved for operations.

The three-layer PGPE model provides abroad guide for the evidencing process. The


P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S xi

5. For example, most CGAC pilots combined a background paper on country-level issues with specific papers on various sectorsor themes.

first layer, that is, the problem to be analyzed, isusually what motivates the analysis. The analyticdescription of this layer would be drawn byteams, as a first step, from existing insights or ex-isting materials. At the heart of the analysis arelayers two and three: institutional structures andgovernance arrangements, and stakeholders,their interests and incentives, and the way theseinteract with formal and informal institutions.

For these layers, the analysis can draw on arange of sources. Generically speaking, thesources are as follows:

• Pre-existing diagnostics (such as sector ESW,PERs, CPARs, and PETS),

• Laws and regulations, • Organigrams and mapping of de jure and ac-

tual process flows as important sources tocapture institutional structures,

• Sector-specific data (e.g. pricing informa-tion, bill collection),

• Public opinion surveys (existing or commis-sioned), and

• Other written sources, such as media cover-age of policy debates and minutes of parlia-mentary debates.

Interviews with individuals and with focusgroups, where appropriate, are also a key sourceof information. Interviewees may includeknowledgeable local stakeholders—researchers,journalists, civil society representatives, and pol-icy thinkers within government and politicalparties—as well as in-depth discussions withthe Bank’s country or sector teams to capturetheir often considerable (tacit) knowledge aboutpolitical economy issues.

A core principle for evidencing GPE analy-sis is triangulation. It is helpful if more anec-dotal or ‘soft’ information can be triangulatedwith ‘harder’ sources, such as actual budget allo-

cations, pricing information, asset declarations ofpublic officials, or the tracing of publicly acces-sible sources that reflect positions taken by indi-viduals on the issue at hand (such as electricitytariffs, teacher training and wages, and gover-nance arrangements for the financial sector).

GETTING THE PROCESS RIGHT:ISSUES TO CONSIDER

Getting the process right is essential for PGPEanalysis to be useful and effective. EffectivePGPE analysis benefits from a clear vision of theintention of the work—that is, whether it ismeant only to inform Bank teams internally orto inform wider debates, and from a clear visionof the focus and the type of output that is beingsought. Furthermore, it is vital to consider thefollow-up process concerning how the analysiswill feed into relevant programming, strategies,and operations. This paper discusses five processissues:

(i) Planning and timing of PGPE work(ii) Defining and finding the necessary skills(iii) Implementing the diagnostics(iv) Sharing and disseminating outputs(v) Bridging analysis and follow-up/action.

All five of these issues are important for theproduction of good PGPE diagnostics. More-over, making the link between the diagnosticphase and the follow-up phase—including feed-ing into strategies and operations—deservesgreat attention.

KEY CONCEPTS

Annex 1 discusses selected key concepts and ap-proaches that are considered useful for framingapplied PGPE diagnostics. Several of these con-


xii P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

cepts are widely used, by DfID’s Drivers ofChange approaches and for the EU’s sector gov-ernance assessments.

Three clusters of drivers: structures,institutions, stakeholder (actors)

Three key types of factors or variables are com-monly considered in PGPE analysis: structural,institutional, and actors. Structural factors are be-yond the direct control of (local) stakeholders,and many such factors change only slowly overtime, although some, such as commodity prices,can be highly volatile. Institutional variables arethose related to ‘the rules of the game’ (lawsand regulations, as well as informal rules such associal obligations). Actors or stakeholders com-prise individuals as well as organized groups orgroups with shared interests, such as politicalparties, the military (and in some cases, power-ful secret services), business associations, non-governmental organizations (NGOs), traditionalassociations, and traders in a particular region.A crucial task of PGPE analysis is to capture howstructural, institutional, and stakeholder vari-ables interact and how they impact on the spe-cific development challenges at hand.

Capturing historical legacies

Historical legacies often have profound effectson shaping current dynamics. Capturing suchlonger processes, and how societies continue todeal with them, provides depth and perspectiveto the issue of ‘how things have become theway they are today’. While some countries’ lega-cies are well analyzed and understood, manyWorld Bank teams work in countries that havereceived limited attention from academic schol-ars, or where existing attention has never reallyexplored the link between legacies and present

constellations. In particular, past conflicts andtensions may be important to understand asthey can shape attitudes long beyond their orig-inal occurrence.

Mapping formal and informal institutions

Mapping institutions is valuable in two ways:First, it creates a clearer picture about the settingin which stakeholders operate, and how thisshapes their interests and incentives, and helpsin developing a more detailed understanding ofthe client agencies with which Bank programsinteract. Second, institutional mapping is valu-able to identify potential levers of change. In-stitutional mapping comprises an analysis offormal as well as informal rules, and the ways inwhich they are related. In many client countriesthis relationship is complex, formal rules beingpresent but being observed only selectively whileinformal rules are influential.

Stakeholder analysis

Stakeholders come in many guises. They can beindividuals or they can be specific groups, suchas the mid-level officials of a ministry, businessor farmers’ associations, or political parties.Three main ways of categorizing stakeholdershave become widely used: ‘demand-side’ versus‘supply-side’ actors/stakeholders; reform cham-pions versus reform opponents; and ‘winners’,‘neutrals/undecideds’ and ‘losers’. Such catego-rizations can be useful as long as they do not losesight of ambiguity and dynamics.

‘Reform champions’ are particularly impor-tant for promoting change. However, individualchampions often have multiple agendas and is-sues that they need to take into account, such asmaintaining their own position and/or powerbases, possibly addressing a range of challenges


P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S xiii

that may require horse-trading and compro-mising with others, having important personalinterests or obligations, and so on. Also, some re-form champions may be strong individual sup-porters of reform but be poor at coalition build-ing, which can limit their effectiveness.

Resources, rent distribution, politicalstability, and legitimacy

Understanding the political economy of rents,how it is linked to the distribution of power insociety, and how this interaction in turn affectsgrowth, poverty reduction, and reform processesis a critical part of problem-driven analysis. Theallocation of rents and how that evolves overtime are crucial for understanding underlyingpolitical- economy drivers, especially at thecountry level, but often also at the sector and

policy level.6 One challenge is that there areseveral different diagnostic as well as normativeperspectives on rents. While a more traditionalperspective holds that rents and rent-seekingshould be minimized, Khan and Jomo (2000),Haber et al. (2003), and others caution (a) thatsome regulatory rents can be important for eco-nomic development, and (b) that the distribu-tion of rents is frequently an important pillar ofpolitical stability and that attempts to curtail orredistribute rents can result in political instabil-ity. Overall, the nature and allocation of rentscan be harmful, stifling the economy and stok-ing discontent, or it can be relatively benign,promoting development and based on an im-plicit social contract. However, it can be chal-lenging to judge on an ongoing basis whichequilibrium a country or a sector is in, given thatthe boundaries can be fuzzy.


xiv P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

6. The standard definition of economic rents refers to the difference between what a factor of production is paid and how muchit would need to be paid to remain in its current use. There are different types of rents, most importantly those linked to con-trol over natural resources, those derived from government regulation, and those related to illicit activities such as narcoticssmuggling.

P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S xv

AAA Analytic and Advisory ActivitiesAC Anti-CorruptionCPAR Country Procurement Assessment

ReportCAS Country Assistance StrategyCD Country DirectorCDD Community Driven DevelopmentCEM Country Economic MemorandumCM Country ManagerCMU Country Management UnitCPC Country Program CoordinatorCPS Country Partnership StrategyCSC Country Sector CoordinatorDPL Development Policy LoanESW Economic and Sector WorkFY Fiscal Year

GAC Governance and Anti-CorruptionGPE Governance and Political EconomyIEG Independent Evaluation GroupIGR Institutional and Governance ReviewISN Interim Strategy NoteMDB Multilateral Development BankPA Poverty AssessmentPER Public Expenditure ReviewPETS Public Expenditure Tracking SurveyPGPE Problem-Driven Governance and

Political Economy PSIA Poverty and Social Impact AnalysisSOE State Owned EnterpriseSTC Short Term ConsultantTA Technical AssistanceTTL Task Team Leader

Abbreviations

Introduction: Why Governanceand Political Economy Analysis?

P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

Politics and the political economy matter forwhether and how reforms happen—in devel-oping as well as in developed countries. TheWorld Bank as an institution and its individualcountry and task teams have been grapplingwith this issue for many years.1 A number ofteams have experimented with various ways ofanalyzing and understanding the political econ-omy context of reforms using such diagnosticsfor smarter engagement. The good practiceframework described here is an attempt to sum-marize relevant analytic tools and approaches, toindicate how they can be used (more) systemat-ically, and to make key lessons readily available.It also seeks to set out how such tools can beused in a way that is problem driven, that is, fo-cused on specific issues and challenges ratherthan on developing broad overviews, in order togenerate operationally relevant findings and im-plications.

Integrating governance and political econ-omy analysis more systematically into Bankoperational work is important to enhance de-velopment effectiveness, to better address risks,and to respond to client demands for ap-

proaches that are tailored to specific situations.A number of recent evaluations have under-lined the need for understanding the politicaleconomy context of reforms more systemati-cally and for taking this into account when de-signing and implementing reforms (for example,IEG 2006, 2008). One client survey in a majorcountry noted that ‘clients considered the Bank’sgreatest weakness to be its lack of considerationfor political realities on the ground and, broadlyspeaking, conducting business in a bureaucraticway that is not attuned to country conditions’(World Bank 2008c: 15). GPE analysis can helpto anticipate and manage risks—including risksof reform failure, of Bank-supported reformstriggering unintended negative consequences,as well as potential reputational risks. It can alsoassist in transmitting important knowledgeabout institutions and stakeholders more quicklyand effectively to staff newly joining a countryor other operational team.

GPE analysis will not lead to quick fixes,but it can be practical and useful for enhanc-ing strategies and operations. Understandingwhat motivates stakeholders, the sources and

1

1. While the World Bank’s mandate explicitly precludes it from engaging in politics, it has become widely accepted that the in-stitution needs to understand the political economy context of reforms from a diagnostic perspective, in order to be able toassist countries effectively in designing and implementing development strategies and policies.

distribution of rents, or the interaction of formaland informal institutions cannot always be eas-ily translated into building pro-reform coali-tions or finding quick fixes to re-aligning in-centive structures. Thus, GPE analysis shouldnot be expected to deliver a new magic bullet. Atthe same time, recent experience shows thatGPE analysis can be practically oriented andvaluable and can set out options and solutionsrather than mainly pointing to obstacles. There-fore, if used wisely, high quality GPE analysiscan certainly add to the way the Bank does busi-ness and ultimately make a crucial contribu-tion to increasing development effectiveness.

The objective of this framework is to sys-tematize approaches to GPE analysis and toprovide orientation for teams that are consider-ing undertaking it. The framework especiallydraws on the experience with a number of pilotstudies undertaken in FY08 and FY09, as well ason earlier studies. These have included country-level analysis as well as analysis focused on a va-riety of sectors—electricity, transport, telecom-munications, water, and public sectorreforms—and on thematic challenges, especiallythe management of natural resource wealth.More broadly, providing a synthesis of these ex-periences and guidance is intended to supportGAC implementation; as well as ultimately tocontribute to the wider goal of strengthening de-velopment effectiveness.

The intention of this framework is to pres-ent a menu of options, rather than to offer orprescribe one particular way of doing things.Such an approach is hoped to be the best way toprovide guidance that is practical and specific,but which can and should be tailored to partic-ular country circumstances and to specific ques-tions to be addressed. The framework does notset out a particular product. GPE analysis cantake a variety of forms, from major free-stand-ing analytic work, such as has been done for

some of the CGAC pilots, to background notes.It has included notes feeding into economicand sector work such as Public Expenditure Re-views (PERs), Poverty Assessments (PAs), orcountry economic memorandums (CEMs), andjust-in-time notes written directly for opera-tional teams during project preparation or im-plementation.

The framework is grounded in the belief thatacross the Bank, teams can benefit significantlyfrom applying GPE analysis more frequentlyand routinely. Furthermore, the framework isprimarily based on a diagnostic rather than aprescriptive (or normative) approach to gover-nance (see also Rodrik 2008a). While the ulti-mate goal is to improve governance and devel-opment outcomes, accomplishing that requirescarefully understanding the existing governancearrangements and their political economy driv-ers. Such understanding then allows the designof more feasible, as well as smarter, innovativepathways to reform, adapted to the specific en-vironment.

The emphasis of this note is on good diag-nostics. This is complemented by a brief dis-cussion of how analysis can be translated intoaction, presented in Part Two. This aspect willbe explored more fully in an accompanying pa-per. A further paper is expected to address theimportant issue of how country teams can con-tinuously and more systematically monitor theevolution of political trends on the ground.

This framework has been developed to fa-cilitate greater attention to and use of GPEanalysis among Bank management, teams,and staff. The framework is therefore tailored tothe context of Bank operations and strategy de-velopment (such as country assistance strate-gies) designed to show how GPE analysis can beused to inform and shape them to support moreeffective development. The focus on Bank strate-gies and operations is not meant to imply that

I N T R O D U C T I O N

2 P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S

these are the most crucial processes at the coun-try or sector level. Rather, such a focus is essen-tial because the purpose of this framework is toencourage and facilitate thinking about contex-tual factors among Bank Group teams. At thesame time, a number of the ideas and discus-sions in this document may also be of interest toothers, such as policy researchers and CSOs inclient countries, other development agencies,and MDBs.

Although only some Bank staff will directlycarry out GPE analysis, many teams andmanagers, across all sectors and levels, canbenefit from understanding the potential useof such work, when and how it can be applied,and lessons learned. Managers, in particular,may want to make sure that teams pay attentionto GPE aspects when developing strategies andoperations. For TTLs and staff, a general un-derstanding is important in order to assess whenan exploration of GPE factors would be neces-sary or beneficial and when staff or consultantswith relevant skills need to be included on teamsand to be able to judge and act on the implica-tions from such diagnostics.

The framework is divided here into five parts:Part One discusses the overall rationale of thisframework and sets out the key foundations. Itproposes that a ‘problem-driven’ approach toGPE analysis (that is, PGPE) is most likely toyield operationally useful insights, and ulti-mately to help enhance development effective-ness. Part Two discusses how PGPE analysis canbe used to inform and shape Bank strategiesand operations, and offers options for translat-ing the analysis into action more broadly. Theserange from a reactive approach of adjustingstrategies or operations to the existing space for

reforms to a proactive approach that seeks to ex-pand this space.

Part Three sets out the various levels of analy-sis that may be undertaken, from an overallcountry focus, to a sector or thematic focus, toa GPE analysis that is undertaken to informspecific projects or policy decisions. PGPEanalysis is valuable at each of these levels. Ideally,teams should ensure that some analysis of coun-try-level drivers and dynamics is available beforedrilling down into more specific areas of con-cern. In turn, such drilling down is essential tomaximize operational usefulness. Part Four ad-dresses the issue of evidencing a GPE analysis, akey challenge in producing high-quality work.Part Five addresses process issues that arise whenundertaking GPE-type analysis. GPE analysis isstill a relatively new diagnostic approach at theWorld Bank, which raises challenges in definingthe work to be done, finding the right skill set,and managing dissemination. GPE analysis willoften raise sensitive issues, so some up-frontguidance, drawing on the experience of pilotteams, may be useful.

Annex 1 covers key political economy con-cepts and approaches. It provides a brief sum-mary of the concepts underlying the questionsthat PGPE analysis can address. Annex 2 pro-vides a more detailed discussion of analysis atcountry, sector/thematic, and project levels. Itprovides an overview of how the general princi-ples, approaches, and options can be appliedacross the levels of analysis, and points to someof the specifics concerning various levels (such asthe varying challenges in evidencing the analy-sis). Annex 3 reflects the specific questions forcountry-level analysis that were originally de-veloped in FY07-8 for the CGAC pilot studies.

W H Y G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S ?

P R O B L E M - D R I V E N G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y A N A LYS I S 3

The emphasis of this framework is on prob-lem-driven analysis to address developmentconstraints. The framework emphasizes GPEanalysis that focuses on particular problems,challenges, or opportunities—such as addressinga failure of existing policies to yield tangible re-sults in poverty reduction, managing a resourceboom-and-bust, or, at the sector level, analyzingwhy power or health sector reforms have re-peatedly stalled or have failed and what could bedone differently to move forward. Importantly,‘problem-driven’ does not mean focusing exclu-sively on areas of difficulty. For example, in arange of environments much can be learnedfrom an analysis of how islands of excellencehave emerged. In other cases, the key issue to an-alyze may be how to react to a window of op-portunity that appears to be opening up. ’Prob-lem-driven’ therefore means focusing on specificquestions and challenges, in contrast to other

analysis that provides broad overviews, eithergenerally or against certain benchmarks.1

The use of various types of GPE analysis toassess risks and to help shape reforms is be-coming increasingly widespread across theprivate and public sectors. The private sectoris using progressively more sophisticated politi-cal risk assessments.2 The OECD is adopting aprogram on ‘Making Reform Happen’, intendedto support member countries in better analyzingpolitical economy factors in reforms and to as-sist them in building successful advocacy andcoalitions for reform.3

More and more international and aid agen-cies regard having a good understanding of thecontext of their operations as an essential part oftheir business. The most well-known approachis DfID’s Drivers of Change, followed by thepublication in 2009 of a ‘How to Note’ on po-litical economy analysis.4 The EU, the Nether-


Part One: A Problem-DrivenApproach to Governance

and Political Economy

1. These are useful for mapping where countries stand relative to others—and in fact PGPE analysis may draw on such assess-ments. However, broad overviews often do not provide clear guidance or implications for specific strategic or operationalquestions.

2. See e.g. Pricewaterhouse Coopers: http://www.pwc.com/extweb/pwcpublications.nsf/docid/EAB01AC994713716852570FF006868B6.

3. See OECD http://www.oecd.org/document/15/0,3343,en_2649_34487_41399503_1_1_1_1,00.html.4. DfID: http://www.dfid.gov.uk/aboutDFID/organisation/driversofchange.asp; http://www. gsdrc.org/go/topic-guides/

political-economy-analysis/tools-for-political-economy-analysis.

lands, SIDA, and CIDA are also experimentingwith various assessment frameworks.5 Theseframeworks used by other donors to date havelargely been focused on the country level only(rather than focusing also on specific sectors orissues) and have generally not been problem-driven but rather provided general overviews.There is currently a growing interest amongseveral of these agencies to develop more sector-focused approaches to governance and politicaleconomy analysis, since this is seen as the wayforward.

Around the World Bank, GPE-type diag-nostics have been carried out by various teamsfor a number of years. Two of the key aims ofthis framework are to make the ideas and ap-proaches explored by those teams more read-ily available and to promote the emergence of

more systematic good practices. Existing ex-amples of governance and/or political economyanalyses include: (i) institutional and gover-nance reviews (IGRs6) (ii) governance diagnos-tic surveys, focused on understanding corrup-tion overall and on the specific interactionsbetween government and citizens and/or theprivate sector,7 (iii) work focused on the distri-butional dimensions of policy reforms and onpolitical economy factors related to these as partof poverty and social impact analyses (PSIAs)8,and (iv) a wide variety of GPE analysis done asbackground notes for country assistance strate-gies (CAS) and/or as part of CGAC processes,feeding into major economic and sector work(ESW) (such as country economic memorandaor PAs), and notes on GPE issues in resourcerich environments and regarding food security,

PA RT O N E


B o x 1 Using governance indicators and governance surveys as inputs to GPE analysis

A related but distinct area of work is that aimed at producing governance indicators, ‘governanceat a glance’ summaries and comparative charts of indicators, and governance diagnostic surveys.Governance indicators can be very useful to provide a broad assessment of where governancestands in a country, especially in comparison to regional or income-level peer groups. However,they do not provide the detail and depth of analysis needed to define operational implications.Efforts to develop more actionable governance indicators (AGIs) for various areas are under way.a

Governance indicators can be very useful as an input to GPE-type analysis. Importantly, there arestill methodological problems about many governance indicators—so some caution in theirinterpretation is required. Furthermore, most such indicators do not provide information onunderlying political economy drivers.

Source: Authors.a. See: http://go.worldbank.org/16E7O0VXW0 (World Bank intranet only).

5. Netherlands: http://www.clingendael.nl/publications/2007/20071000_cru_occ_unsworth.pdf. In early 2008, the OECD organized a Conference on Governance Assessments, the documentation for which includes use-

ful material and links to governance assessment tools.6. http://intranet.worldbank.org/WBSITE/INTRANET/UNITS/INTPREMNET/0,,contentMDK:20407733~pagePK:

64156298~piPK:64152276~theSitePK:489837,00.html. 7. See http://go.worldbank.org/KK9UICJUQ0 (World Bank intranet only). Another website brings together all the questions

that have been used in different surveys related to a range of sectors and governance issues (e.g. corruption, nepotism, re-sponsiveness: http://go.worldbank.org/KS54ICJMM0 (World Bank intranet only).

8. See: http://go.worldbank.org/39I9SFVEJ0. The Social Development (SDV) anchor within the Sustainable Development Net-work (SDN) has recently provided a framework that summarizes the experience with political economy analysis as part ofPoverty and Social Impact Analysis (see World Bank 2008b and 2007c).

as well as GPE notes done to inform specificprojects.9 This framework draws in particular onthe experiences generated by these various GPEbackground notes. Such notes have often notbeen shared across teams, which has hamperedthe building up of good practices.

This framework builds on these existingapproaches and pilots. Drawing on the lessonslearned, it emphasizes a focus on specificproblems or vulnerabilities, as well as theneed to understand political economy driversby examining them in a systematic way. Theseelements contribute to making GPE analysesoperationally useable, so that they can con-tribute to development effectiveness. The frame-work also proposes some clearer standards andexpectations, while keeping in mind that thestandards have to be realistic regarding whatcan be done under real-world conditions andwith finite resources.

I.1. THREE STEPS OF PROBLEM-DRIVEN GPE ANALYSIS

A helpful way to think about a problem-dri-ven approach to governance and politicaleconomy analysis is to distinguish three layers:(i) identifying the problem, issue, or vulnera-bility to be addressed, (ii) mapping out the in-stitutional and governance arrangements andweaknesses, and (iii) drilling down to the polit-ical economy drivers, both in identifying obsta-cles to progressive change, and in understandingwhere potentially a ‘drive’ for positive changecould be emerging from (see Figure 1). This ba-sic approach can be applied to analysis at thecountry, sector, or project levels—with appro-priate adaptation and tailoring as discussed inPart Three (and further elaborated in Annex II).

As reflected in Figure 1, as a first step it isimportant to define the challenge or problem

A P R O B L E M - D R I V E N A P P R OAC H TO G O V E R N A N C E A N D P O L I T I C A L E C O N O M Y


F i g u re 1 : Three layers of problem-driven GPE analysis

Whatvulnerabilities/

challenges?

Evidence of pooroutcomes towhich GPEweaknessesappear tocontribute

E.g.: repeated failure to adopt sector reforms.Poor sector outcomes. Infrastructure identifiedas constraint to growth but not effectivelyaddressed. Continuous food insecurity.Corruption continues to undermine the businessclimate even after anti-corruption law

GPE

an

alys

is

Institutional &governancearrangements& capacities

What are theassociatedinstitutional set-up andgovernance arrangements?

Mapping of relevant branches of goverment,ministries, agencies, and SOEs and theirinteraction. Existing laws and regulations. Policyprocesses (formal rules and de facto). Whatmechanisms intended to ensure integrity andaccountability and to limit corruption exist?

Politicaleconomydrivers

Why are thingsthis way? Whyare policies orinstitutionsarrangements notbeing improved?

Analysis of stakeholders, incentives, rents/rent distribution, historical legacies, and priorexperiences with reforms, social trends andforces (e.g., ethnic tensions), and how theyshape current stakeholder positions and actions

Pro

ble

m-d

riv

en

9. Most of these types of analysis have been carried out on a pilot basis, with a few cases studied for each type.

and establish that it appears to have a gover-nance or political economy dimension.10 Theinitial identification of challenges or problemsmay emerge from core Bank work, such as eco-nomic and sector work and country assistancestrategy completion and progress reports, orfrom the ongoing policy dialogue with the gov-ernment or other stakeholders. In a forward-looking way, the team may want to assess howchallenging the reforms they are planning tosupport are likely to be, given the institutionaland stakeholder environments. They may alsowant to anticipate governance and politicaleconomy risks in operations such as major pol-icy or investment lending.

In many cases, this is not about identifyinggovernance as the main overall bottleneck todevelopment. Rather, it means consideringwhether other development challenges—suchas improving infrastructure or getting tractionon pension systems reform or managing urbandevelopment—have governance and politicaleconomy dimensions to them that need to beaddressed. Typical challenges are reforms thatfail or are only partially implemented, as well asreforms that are undertaken but have significantnegative and unexpected results. More macro-level problems may include a consistent failureto achieve significant poverty reduction, possi-bly despite substantial growth. Such challengeshave been experienced across client countriesand areas of Bank operations. They have oc-curred in agriculture, health, and public ad-ministration reforms, as well as in fundamentalpolicy commitments inscribed in national de-velopment strategies.

The second and third layer of PGPE analy-sis aim at understanding the institutionaland governance arrangements and how these

are related to poor outcomes, and then atdrilling down to the underlying political econ-omy factors and drivers. The aim is to explainwhy policies and/or institutional and gover-nance arrangements are not sufficiently sup-portive of the development issue of concern. Theexplanation will involve an analysis of stake-holders and their interests and incentives, howthese interact with the institutional environ-ment (both formal and informal), and how thesehave been shaped by country or sector dynam-ics over time. Figure 4 in Annex 1 maps out theinteractions between stakeholders, institutions,and structural factors. Part Four provides somebasic principles regarding how such analyses canbe evidenced to be credible.

Layers two and three, as presented here,clearly overlap. That is, an analysis of insti-tutions is widely considered part and parcel ofany political economy analysis. The two lay-ers are nonetheless differentiated for two rea-sons. On the one hand, many types of existinggovernance analysis include layer two, but ex-clude layer three; they identify governance weak-nesses, but do not analyze why they exist. Onthe other hand, some political economy analy-sis is predominantly focused on an analysis ofstakeholders. However, understanding the in-stitutional status quo in sufficient detail is cru-cial not only for GPE analysis as such, but alsoin order to be able to map a feasible reformpath. Thorough institutional analysis strength-ens the operational usefulness of GPE analysisby identifying what reforms are feasible from apolitical economy as well as an explicitly insti-tutional perspective (the degree of detail inwhich analysis is carried out of course varies, de-pending on a team’s needs and the time avail-able).

PA RT O N E


10. This is essential for establishing that governance creates a problem/hinders development, and the nature of the problem. Thiscreates a clear focus and makes the analysis problem-driven, rather than jumping directly to a discussion of governance weak-nesses and problems, of which there are many in most contexts.

The important goal of drilling down to thepolitical economy layer of analysis is to under-stand why the problem that has been identifiedis not being addressed successfully, whether theeffort has been to change governance and insti-tutional arrangements or to adopt better poli-cies. For a planned operation, the political econ-

omy layer can assess the relative likelihood ofstakeholder opposition or of failure to mitigatepolicy or governance risks successfully. Politicaleconomy analysis asks about underlying drivers,such as the relationship between stakeholders,available rents and how they are distributed, in-terests, collective action dilemmas, and incen-



B o x 2 : Defining feasible solutions—Zambia and Mongolia

In Mongolia, domestic funding for public investments increased more than fivefold between theearly 2000s and 2008; but public investment planning remained rather ad hoc and subject topolitical jockeying—during budget preparation as well as at parliamentary approval stages—withlimited regard for social and economic priorities.

One feature of the situation has been that the senior civil servants and politicians who expressthe greatest frustration with the existing situation favor the (re-)creation of a ministry or agencyof planning. This is not considered to be good international practice, as there is a perceived riskof creating unproductive rivalries between the Ministry of Finance and such an agency or ministry.However, a planning agency is appealing to local stakeholders, since a similar institution existedin the past, which is seen as having delivered better results than the current system.

Thinking about feasible options suggests, first, that the emphasis should be on working withthose interested in improvements—rather than insisting on particular institutional models. Second,it may not be realistic to de-politicize public investments overall, but rather it may be realistic tofocus on some ‘bright lines’—such as simple, clear rules restricting the ad hoc adding of projectsto public investment plan (PIP) lists, and promoting the vetting of major projects, which could helpto limit the space for politicization and ensure a better targeting of funds on agreed priorities.

The Zambia ‘GAC for CAS’a pilot took a close look at governance and political economy issuesconcerning the power sector. The experience over the past decade has been one of ambitiousefforts at policy reforms (unbundling, tariff restructuring and increases, creation of public-privatepartnerships, creation of an impartial regulator), followed by reversals when the governmentpulled back from unbundling and privatization plans. At the same time, the need for expansionand better maintenance in the electricity sector has intensified in the context of a growingeconomy and a booming mining sector.

The diagnostic analyzes the important role that the state-owned electricity company ZESCOplays in the context of Zambia’s patronage networks. Furthermore, mines and (better off) urbanconsumers have been the main beneficiaries of low electricity tariffs, while most of the poor haveremained unconnected.

A potential feasible solution involves making ZESCO a full participant in planning reforms(rather than seeking its dismantling). Furthermore, an approach focused on feasibility couldinvolve prioritizing to add new-generation capacity on the basis of full-cost pricing for theincrement. Stakeholders with the greatest interest in better supply—mining companies and otherpotential business users—would pay full-cost tariffs for the added electricity, while investmentswould not be held back by waiting for overall tariff reform to materialize. For the medium term,the proposal is to engage private business associations and consumer associations in electricity-sector expansion and reform, in order to increase demand and to break out of the low-levelequilibrium of unwillingness to pay higher tariffs and shoddy utilities.

Sources: Mongolia: The Political Economy of the Resource Paradox (2009); Zambia: Governance, Political Econ-omy and Development Strategy (2007).a. ‘GAC for CAS’ was the term adopted for the country-level governance and political economy analyses un-dertaken in FY08 to inform upcoming country assistance strategies and as part of the overall governance andanti-corruption (GAC) implementation.

tives. Understanding such underlying factorsbetter is important in efforts to improve policiesand governance arrangements and ultimately inachieving development results.

In summary, a key lesson from experienceis that GPE analysis is most useful when it ismotivated to understand specific issues orchallenges and hence is ‘problem-driven’. Fur-thermore, it is important to combine gover-nance and institutional analysis with an analy-sis of underlying political economy drivers. Inthis way, GPE analysis goes beyond most typesof governance assessments, which focus on iden-tifying governance weaknesses but provide littleexploration of why these exist or of the dynam-ics that could lead to change. At the same time,it emphasizes an explicit coverage of institu-tions rather than of stakeholders only.

The next section briefly sets out in whattypes of situations governance and politicaleconomy analysis as laid out here is most crucialand valuable. Section 1.3 discusses the crucialaim of PGPE analysis, namely to help teamsidentify feasible approaches to reforms. PartTwo then turns to the practical questions ofhow GPE analysis can be translated into actionor ‘doing things differently’. Parts Three andFour deepen the discussion about the analysis assuch, the range of issues it may focus on, andhow it can be evidenced.

I.2. WHEN IS GOVERNANCE ANDPOLITICAL ECONOMY ANALYSISMOST CRUCIAL AND VALUABLE?

Governance and political economy analysis canbe valuable in a variety of different contexts andsituations. First and foremost, GPE analysis isimportant where and when governance and po-

litical economy factors appear to preventprogress that is otherwise considered possiblefrom a technical perspective (‘developmentalrisk’). Secondly, country-level GPE analysis canbe particularly relevant also in situations wheresignificant reputational risks for the World Bankexist. Thirdly, where fiduciary risks are high,GPE can be valuable for understanding the driv-ers of a poor fiduciary environment and to de-fine adequate responses.11

As Part Three below sets out, GPE analysiscan be targeted at (i) the country, (ii) the sectoror thematic or (iii) the particular policy or proj-ect level. Seeking to understand country-leveldynamics is especially relevant duringCAS/CPS/ISN preparation periods. This is par-ticularly important when there is a sense that theoverall country program or significant parts ofthe policy dialogue have not borne fruit in thepast (e.g. as documented in the CAS ProgressReport) or there have been a number of surprisesin terms of policy action or inaction, or ex-pected development progress that failed to ma-terialize. It can also be important in situationswhen significant change has taken place in acountry—such as a change in government, theemergence of natural resource exploitation, etc.Furthermore, country-level analysis is crucial insituations where major governance-related risksexist that could cause a halt to the Bank’s over-all country program (e.g. risks of state fragility,a coup d’état, etc.), or where other types of rep-utational risks for the Bank are present. Finally,country-level GPE analysis provides an impor-tant ‘public good’ that benefits all the variousteams working on a country—by facilitating aquicker, better understanding of country contextwhen team members rotate, as well as by pro-viding important background material if and

PA RT O N E


11. On these three types of risks see World Bank (2007a) (‘GAC Strategy’).

when the need for a more detailed understand-ing of political economy drivers emerges in a sec-tor or related to particular projects.

Sector-level GPE analysis is especially relevantwhen envisaged reforms in a sector are signifi-cant and are likely to be difficult because oftheir distributional impact, their institutionalcomplexity, and/or the structure of incentivesand constraints. It is also important wherechanges in a sector or policy area have been onthe agenda for a long time without ever materi-alizing; or when earlier improvements in sectorpolicies or outcomes were reversed. A relatedconsideration is the ability of governments to‘make reforms happen’, i.e. to manage institu-tional change as well as different stakeholderinterests productively. Especially where experi-ence suggests that this ability is limited andwhere consequently reforms may be left unad-dressed, blocked or mismanaged, it can be cru-cial for a Bank team to invest in GPE-typeanalysis in order to gain a better understandingof the drivers and dynamics at play and how theteam might best structure its own interventionsand support. Ideally, sector-level analysis canbuild and draw on existing country-level work,which will signal overarching constraints and dy-namics that can affect most or all sectors. Fur-thermore, often it can be helpful to use a value-chain approach for structuring sector as well asproject focused analysis (see Campos and Prad-han 2007) and as a way to identify key points ofrisk or weakness in a sector.

For specific projects or policy questions (suchas triggers to be used for policy lending), GPEanalysis is most crucial if significant governance-related problems—including but not only cor-ruption risks—are anticipated. In some con-texts, a brief GPE diagnostic will be valuable inorder to develop meaningful risk matrices, in-cluding a better understanding of the GPE driv-ers underlying fiduciary risks. In other cases,

e.g. CDD-type projects or projects that aim toachieve positive distributional impacts for poorergroups, some understanding of local politicaleconomy drivers and risks of capture are essen-tial for designing effective interventions andmonitoring frameworks. Generally, many of thereforms that project teams seek to achieve—amore meritocratic civil service, better financialsector regulation, more effective management ofnatural resources, or better access to social serv-ices for the poor—involve changes to power re-lations, rent-seeking opportunities, or to in-grained incentive structures. Consequently,paying attention to GPE related risks and to pos-sible mitigation and management strategies isimportant.

The scope and depth of PGPE-type analysiscan be adjusted to fit what is needed in a par-ticular case. Often, there are opportunities todraw on existing analysis—for example througha workshop with relevant local or internationalresearchers. However, if the need is for more tar-geted information and answers to specific ques-tions—such as seeking a GPE perspective on thedesign of CAS pillars, or to understand whypower sector reforms have not materialized de-spite substantial policy dialogue and investmentlending—and for advice on implications forBank interventions, undertaking or commis-sioning dedicated analytic work is essential;while there are a range of options for fittingthis within available resource and time con-straints (see also Part Five).

A logical and coherent way for developingGPE analysis for sectors and projects/specificpolicies is by building on prior analysis of coun-try-level dynamics (see also Annex 2 for a morein-depth discussion of the three levels). For anincreasing number of countries, such analysis isnow available, either undertaken by World Bankteams or other development partners. If theneed for project/policy focused GPE diagnostics



emerges where no prior analysis of country andsector level dynamics is available, it may often benecessary to include at least a brief sketch ofcountry-level dynamics as well as of the overallvalue-chain of the sector. Some project-relatedproblems are likely to be rooted in the widercountry context (e.g. bribes and kick-backslinked to country-level problems with illicit fi-nancing of political election campaigns), otherscan be linked to other ‘break’ points in the sec-tor value chain—further upstream or down-stream—while yet others are highly specific tothe particular project intervention or locality.Therefore, in order to generate sensible solutionsit is important that more micro-level GPE analy-sis is able to link up to wider sector or country-level issues to the degree that this is feasiblegiven existing work, as well as time and resourceconstraints.

The sensitivity of GPE analysis varies signif-icantly across countries and issues. Crucially,and as set out repeatedly in this framework, theWorld Bank undertakes such analysis with theaim of enhancing development effectiveness—including the definition of feasible reforms asdiscussed in the next section. The analytic workmay be done to inform Bank staff and manage-ment only or it may be shared and discussedwith country stakeholders—in whatever form issensible and appropriate to the specific contextand to the overall goal of furthering develop-ment effectiveness (see Part Five for a more de-tailed discussion).

I.2. GOOD ENOUGH GOVERNANCEAND FEASIBLE APPROACHES TOREFORM

In many cases, the implications resulting fromGPE analysis will entail considering feasibleoptions for reforms. As Rodrik (2008b) argues:“…dealing with the institutional landscape in

developing economies requires a second-bestmindset. In such settings, a focus on best-prac-tice institutions not only creates blind spots,leading us to overlook reforms that mightachieve the desired ends at lower cost, it can alsobackfire.” While ‘first best solutions’ are techni-cally superior in a textbook sense, this advantagecan be lost or even reversed under real-worldconditions if the solutions are only partially im-plemented, are stalled or reversed because theyare politically unpalatable, or trigger unintendedconsequences during implementation (such asprivatization creating monopolies or reformsleading to institutional breakdowns rather thanimprovements).

At the same time, ‘feasible’ reform optionsare also clearly distinct from ‘default options’,that is, the reform path or non-reform paththat would result from no intervention or ef-fort at change. Because ‘feasible solutions’ aresensitive to the political economy context, theyshould in principle also be more compatiblewith country ownership, a key principle of thedevelopment effectiveness agenda. It can bechallenging to identify a ‘good enough’ solutionthat is feasible, has a relevant development pay-off, and offers the potential of preparing theground for further improvements later on. How-ever, experience from pilot cases shows thatsearching for such solutions is a worthwhile ex-ercise. It helps teams to think through their as-sumptions and to consider in much greater de-tail the range of short- to medium-term reformoptions. It also helps teams think about thesteps that would be involved in moving from anexisting situation to a better one, such as im-proved regulation or a better distribution of re-sponsibilities or incrementally improved sys-tems.

Finding feasible approaches to reform mayinclude prioritizing what vulnerabilities/con-cerns can be addressed with a reasonable chance

PA RT O N E


of success and proposing how governancearrangements can be improved in a way that (a)is feasible, (b) is not likely to be subverted by po-litical economy drivers, and (c) would channelpolitical economy drivers more productively inachieving development and poverty reduction.This includes interventions, based on a carefulassessment, that might proactively seek to easepolitical economy constraints, for example bysupporting coalitions for change, promoting abetter-informed public debate, and so on.

‘Good enough governance’ captures the ideaof considering the change process as movingfrom worse to better governance environments,and of focusing on priorities and gradual im-provements rather than trying to reform gover-nance wholesale—whether at the macro, core, or

sector level (see Grindle 2007, Khan 2006). Im-portantly, the concept of ‘good enough’ doesnot imply giving up on ‘good governance’ as aprinciple. Rather, it acknowledges the practicalchallenge that under real-world conditions, notall desirable governance improvements are af-fordable or feasible in the short to medium term,and that feasible, gradual improvements mayhave tangible payoffs for development goals.

PGPE analysis can incorporate the notions of‘good enough governance’ and feasible optionsinto the process of defining implications andproposals regarding how best to move forwardwith reforms. As discussed in Part Five, this willbe most useful and operationally relevant if donein close interaction with the country or taskteams working on the reform area concerned.



In the past, political economy analysis has some-times been seen as an approach that is morelikely to point to obstacles than to find solutions.This good practice framework seeks to lay outhow problem-driven GPE diagnostics can ex-plain why reforms may be difficult as well asproviding options for solutions. At the oppositeextreme is a perception that, once we under-stand the interests and incentives of stakehold-ers better, we can manage reforms much moreeffectively, work with champions and coalitionsfor change, foster better leaders, and achieve en-hanced results rather quickly. Such expectationscan also be misleading, not least given the factthat the World Bank is only one among a mul-titude of stakeholders in any given country.Nonetheless, GPE analysis can yield tangible in-cremental benefits for country programs andoperations, and ultimately, therefore, also foraid effectiveness and for a country’s develop-ment outcomes.

Problem-driven GPE analysis can be trans-lated into action in various ways. It can providepolitical intelligence for agreement on what isfeasible within teams and with management;moreover, it can help to shape strategies and op-

erations in ways that range from adjusting themto the existing space for change/reforms, to thedevelopment of proactive strategies for expand-ing the space for change/reforms (see Figure 2).The different ways of using GPE analysis arenot mutually exclusive. For example, country-level analysis may mainly be used to inform acountry strategy, while specific notes may beused to shape a proactive strategy for a particu-lar reform issue, such as a campaign to ensurethat medicines and vaccines are reliably suppliedto rural areas, or to build support for public pro-curement reform. Crucially, in order to provideactionable implications, GPE analysis needs tobe sufficiently specific, and several GPE notesmay be needed if the intent is to inform a rangeof issues and questions. (Note that various op-tions for focusing GPE analysis are discussed inPart Three.) The following paragraphs describethe various uses in greater detail, including ex-amples from recent experience.

Country situations matter for deciding whataction is right. Countries might be in steady-state situations, in which reforms are quite fea-sible but stakeholders feel only limited urgency.They might be in situations of hidden or obvi-


Part Two: Translating Analysisinto Action: Using Problem-

Driven GPE Analysis to EnhanceBank Strategies and Operations

Making reforms happen: translating GPE analysis into action

ous fragility, which often means that manystakeholders are disinclined to take risks. Orthey might be in transformative moments, of-fering wider windows of opportunity as well asrisks and uncertainty (for example over whethermajor shifts in who governs will result in realmomentum for policy improvements). Fiscalwindfalls and, in the reverse, negative fiscalshocks each can create different constraints andopportunities to pursue change. The findingsfrom GPE analysis, especially if it includes thecountry level, as well as the country team’s senseof the situation, can guide the choice of actions.

One important implication of GPE analysiscan be greater awareness of risks and of optionsfor managing risk. This includes a better aware-ness of the potential unintended (negative) con-sequences that first-best solutions may entail.One frequently encountered risk is that of ambi-tious policy reform efforts contributing to signif-icant implementation gaps. In such situations,formal legal provisions are reformed, but incen-

tives, capacity constraints, or the wider institu-tional environment vitiate real implementation,and this in turn can undermine longer-term in-stitutional strengthening as well as the actual re-sults for the sector or issue of concern. Anotherfrequent risk is that of relying strongly on se-lected ‘champions’ of reforms, as well as inter-acting exclusively with the current governmentrather than policy stakeholders more broadly.Such strategies have repeatedly backfired whenchanges in personnel or in the overall govern-ment occur. GPE analysis can indicate optionsfor managing such risks better.

One general lesson on how to translate analy-sis into ‘doing things differently’ is the need forteam and management attention. This includesattention to managing the diagnostic well (asdescribed in detail in Part Five), for example, byintegrating technical and GPE teams whenmajor operations are being prepared. It may re-quire attention to organizing country or sectorteam workshops to discuss findings from GPE

PA RT T W O


F i g u re 2 : Options for translating GPE analysis into action

Selecting operations givenexisting space for change

Seeking to expand spacefor change proactively

Zambia Telecoms: focus on localwinners

Ethiopia Protection of BasicServices (PBS) to mitigatereputational risk: supportsubnational service delivery withparticipation

Mongolia Mining: Technicalassistance (TA) with localthink tank for public debate

Paraguay & BangladeshRoads: external monitoringby stakeholders

The Philippines Publicprocurement reform—proactivecoalition building to combatentrenched corruption networks

⇒ See section 2.1 ⇒ See section 2.2 ⇒ See section 2.3

analyses and to distill what they imply for strate-gies and/or operations; or to call on GPE ex-pertise in the course of implementing anoperation. As described below, if teams decidethat a proactive approach is needed to move re-forms forward, this also requires attention andmanagement. The benefit for teams and man-agement is that ‘GPE-smart’ strategizing mayhave real payoffs for programs and operations.

II.1. INFORMING BANK TEAMS ANDADJUSTING STRATEGIES ANDOPERATIONS TO AN EXISTINGSPACE FOR CHANGE

A basic way of using GPE analysis is to have itinform country strategies and operations (see the

left-hand side of the spectrum in figure 2). GPEanalysis helps to create a better understandingof the environment in which operations takeplace and in which strategic results are beingpursued, and helps to spell out crucial gover-nance and political (economy) risks. This canbe valuable for creating a shared understandingof the environment across a team as well as foragreeing with management and reviewers oncountry programs and/or specific operationsthat are realistic and feasible. GPE analysis alsohelps to preserve and transmit such an under-standing when staff changes.1

Beyond a more limited use of informingteams, GPE analysis should provide and be usedto draw implications for the design of strategiesand operations. In many instances, these impli-

Tr a n s l a t i n g A n a l y s i s i n t o A c t i o n


B o x 3 GAC analysis shaping the CAS—Zambia and Uzbekistan

The Zambia GAC for CAS analysis was carried out from June 2007 to December 2007, in parallelwith the preparation of the 2008–11 CAS, which was adopted in March 2008. The GAC for analy-sis (i) informed the analysis of country conditions alongside other key components, economic con-ditions, and social conditions and poverty reduction, and (ii) informed the World Bank countryprogram focusing on four priority areas: (a) macroeconomic and expenditure management; (b)infrastructure development; (c) institutional capacity enhancement; and (d) social sector develop-ment. Evaluations of earlier CAS periods (1996 to 2003) had shown that progress in Zambia hadbeen unsatisfactory.

The new CAS proposes a filter for any new project proposals, which would, among other con-siderations, clearly focus on feasibility given the governance and political economy context; andwould also promote projects that pursue less far-reaching goals but are more likely to be fully im-plemented (WB 2008: 19–24).

The Uzbek Country Team Learning Event on Governance in CPS (2008) brought together CMUmanagement (CD, CM, CPC), CSCs, TTLs, and the local operational staff to listen to each other andto brainstorm in a structured way about entry points for incorporating governance into the CPSunder challenging political economy conditions. The two-day event was extremely productive: notonly were all those involved required to think about governance as it pertains to their everydaywork, but also as it applies to the Uzbek context. The event culminated in a structured brainstor-ming in small groups that collectively came up with several feasible entry points for governanceengagement in Uzbekistan, which were incorporated into the CPS.

Country examples: Zambia: Brian Levy, PRMPS; Uzbekistan: Jana Kunicova, ECA.\Source: Authors.

1. Often, new staff will need several months to understand the political economy intricacies of their area. Even a brief analysiscan substantially accelerate this understanding, and it can also help to set the political economy problems perceived in a par-ticular area into the wider country context.

cations are ‘reactive’, that is, they are aimed atadjusting strategies and/or operations to an ex-isting space for change:

i. Influencing the country strategy in termsof priorities, and in terms of defining feasi-ble reforms and entry points, as well as pos-sibly proposing new monitoringcomponents that are particularly relevantfrom a political economy perspective.PGPE analysis can be useful at any pointin the country assistance strategy (CAS)cycle, but there is added value if it is doneupstream of the CAS/ISN. At this point, itcan be vital to clarify what the potential forprogress is across the various areas/pillars ofthe CAS, as well as how this envelope maybe stretched. If specific governance-relatedmonitoring indicators are included in theCAS, this helps to keep a focus on these is-sues, even if teams change—in a way that isvery specific and pertinent to the problemsand issues faced by country teams.

ii. Enhancing the design of operations, interms of the what and the how; helping toanticipate problems that are likely toemerge during project implementation andto think ahead about how to manage them.

II.2. ENHANCING AND BROADENINGPOLICY DIALOGUE ANDDEVELOPING INNOVATIVEAPPROACHES TO OPERATIONS

If context permits, GPE analysis can contributeto enhancing the policy dialogue, and to devel-oping innovative approaches to operations.Problem-driven GPE analysis can make impor-tant contributions to enriching the policy dia-logue in-country because it can create a muchbetter understanding of the interests and incen-tives of stakeholders and the formal and infor-

mal institutional landscape within which theyoperate (the middle of the spectrum in figure2). As set out below (Annex 1), a pure ‘commonsense’ approach to political economy can some-times result in perceiving stakeholders as clearlydivided into ‘opponents’ and ‘reform champi-ons’, as well as to a rather narrow focus on cer-tain stakeholders. GPE analysis can foster amore granular understanding of the incentivesand constraints that influence the behavior ofdifferent stakeholders and the implications forpursuing reforms, and can also pro-activelybroaden the circle of stakeholders the Bankseeks to bring together in support of develop-ment policies.

When GPE analysis can be done in an openmanner, it will often involve stakeholders thatmany task teams do not routinely engage with.As a result, it helps establish expanded relation-ships and networks for dialogue and for pro-moting change. A broader set of stakeholdersmay include policy-focused think tanks, policythinkers within political parties (who are notnecessarily members of the executive), andNGO representatives. Moreover, a deepenedunderstanding based on GPE analysis can be animportant boon to working effectively with re-form-oriented individuals and groups/coalitions.

Problem-driven GPE can support innovativeapproaches to operations. For example, teamsmay recognize that it is important as well as fea-sible to promote a better informed public de-bate on reform issues and policy options. Or itmay become apparent that monitoring effortsby local stakeholders can be encouraged andsupported in order to limit corruption andachieve results more effectively. To pursue suchinnovative approaches, a good integration oftechnical and GPE-focused diagnostic work isvital. Management attention and follow-up isalso very important. The mining sector TA inMongolia, described in Box 4, and support for

PA RT T W O


the Paraguay roads project both provide goodexamples of innovative approaches, focusing on(i) promoting a better informed public debateon important issues, and (ii) generating greatertransparency and accountability.

II.3. USING GPE ANALYSIS TODEVELOP CAMPAIGNS FOR CHANGE PROACTIVELY

In a number of instances, the ultimate aim is towiden the space for change (the right-hand sideof the spectrum in figure 2). Various elementsof GPE analysis can help decide whether it isworthwhile to pursue a proactive strategy ofbuilding momentum for reforms. Such a deci-sion is necessarily context- and issue-specific.The main avenues for building traction for

change are coalition-building, information andcommunication campaigns, and generally in-tensified work with stakeholders, in particularstakeholders who can exercise demand for pro-gressive change.

i. Information and communications strategies:such strategies are intended to inform citi-zens at large about the potential benefits ofchange compared to the status quo. Strate-gies would use media and other outlets thatcan reach relevant citizens (such as radiostations broadcasting in rural areas or bill-boards in specific urban neighborhoods).Information campaigns can help govern-ments to move forward with reforms if theyare concerned that the reforms may be un-popular; in other situations, such cam-



B o x 4 : Mongolia—developing an innovative approach to Mining Sector TA

Dealing with a mining boom has been a major policy and political challenge in Mongolia—a coun-try that has rich deposits of copper, gold, coal, uranium, and other minerals. A large number of de-posits have been explored in recent years, while existing mines have become far more profitable.At the same time, there have been numerous public protests over mining sector issues. A key focusis how contracts and agreements are being negotiated with foreign investors. Many Mongoliancitizens broadly favor public ownership of mining; while at the same time also demanding trans-parency from state officials regarding mining sector decisions. This was corroborated by a publicopinion poll undertaken as part of the GPE assessment. The government is keen to benefit fromaccess to international experience and advice on how to tackle the difficult policy challenges. Atthe same time, there has been an element of distrust, as the World Bank is perceived as havingbeen a key supporter of a very liberal mining law that was adopted in 1997 during a period of lowprices. The law has since been revised.

The GPE analysis squarely focused on mining sector issues, covering many of the elements ofthe value chain (see section 3.2), from mining licensing to how the windfall public funds are beingused to foster long-term development. The analysis helped the country team and the task teamthat were preparing a major new Mining Sector TA think through and understand the complexityof the existing institutions and stakeholder interests and incentives, in a way that an exclusivelytechnical analysis would not have been able to. A crucial finding was that there is an urgent needto support a more informed public debate about options for regulating and managing the miningsector and its associated opportunities and risks, and to do so by working with a broad set of sta-keholders, reaching out beyond the executive in multiple ways. In particular, the project will sup-port a local think tank dedicated to mining sector issues in this commodity dependent country.

Source: Authors.

paigns may also be intended to increase de-mand for change, or to focus demand onspecific policy actions.

ii. Intensified work with key stakeholders: forcertain issues of change, it may be crucial tointensify the dialogue with key stakehold-ers, including stakeholders that the Banktraditionally has limited interaction with,such as MPs. Such intensified dialogue canserve to deepen the understanding of re-form issues and/or to win support for cer-tain reform options.

iii. Coalition-building: working pro-activelywith a range of stakeholders interested in

change, including seeking to convince un-decided/hesitant stakeholders to join apro-change platform (see the example ofthe Philippines’ procurement reformprocess, Box 5). Potential pro-reformstakeholders can be brought together invarious forums to develop concrete pro-posals for reform, to agree on positions,and to coordinate actions. The key aimshould be to support progressive coalitions,while accepting that these define the scopeand precise direction of change they wantto pursue (that is, they have ownership ofreform agendas).

PA RT T W O


B o x 5 : Philippines—building a coalition for procurement reform

Public procurement in the Philippines was perceived to be a major governance challenge. A pro-cess of procurement reform was triggered in 1998 by a detailed technical report commissioned byUSAID. Initially, however, the report developed little traction and was shelved. However, a seriesof workshops involving government and nongovernment stakeholders revived demand for pro-curement reform. In early 2001, Procurement Watch (PWI) was established as an advocacy NGO,founded by academics, reputable retired government officials, lawyers, and concerned private sec-tor executives. PWI was funded through an EC-WB grant/trust fund (TF). The grant covered mostof the costs of organizing people and groups in order to advocate for the passage of a new pro-curement bill. PWI raised the awareness for this issue by mobilizing public support and imple-menting a broad media campaign.

In early 2003, a new Government Procurement Reform Act (GPRA) was adopted, a first majorsuccess of the campaign. The challenge since 2003 has been to ensure implementation. Initially, thefocus was on capacity building for government officials at various layers of government. Whilethis was important, after some years there was a sense that advocacy had to be revived in orderto have greater impact. Since then, activities have focused successively on different sectors. Theprocurement of regular goods, for example, became fully transparent and monitorable online[Philippine Government Electronic Procurement System]. The Philippines Boy Scouts became in-volved in monitoring the procurement of school textbooks. Currently, there are plans for seniorcitizens associations to start monitoring drug procurement.

The overarching lessons of this project concerned the importance of getting the right peopleengaged and forming a broad and focused coalition of stakeholders. Proactive and professionalcivil society engagement, the formation of a tightly knit group of reform-minded government of-ficials, the support of progressive legislators who knew how to traverse the complex legislativemaze, the conduct of in-depth technical studies, and the implementation of a well-thought-outcommunication strategy all contributed to effectively mobilizing public action that led to the ad-option of the new procurement legislation and supported its subsequent implementation. Thebiggest challenge remaining is in infrastructure procurement, where only limited but nonethelessnoticeable progress has been made to date.

Source: Ed Campos, Caby Verzosa.

Especially when teams want to use GPEanalysis to pursue a proactive approach tochange, there are several important concerns totake into account. Most crucially, the Bank hasto be an honest broker, and such efforts need tobe in line with the general principle of Bank op-erations to respect the preferences and choices oflocal stakeholders, who will in any case be deci-sive in most instances, especially when it comesto the actual implementation of change.2 There-fore, there is a need for real and substantive di-alogue with a range of local stakeholders.Furthermore, the more an approach is proac-tive, the more it will require resources—includ-ing staff and management time and attention,financial resources to organize meetings, sup-port for local think tanks and NGOs, and fund-ing for communications strategies. At least inpart, the financial resources may be borne byothers (such as government, bilateral donoragencies, and foundations), but at a minimum,Bank teams will need to dedicate staff and man-

agement time. Management attention is partic-ularly important where political sensitivitiesmay need to be addressed at various stages. Thatsaid, the potential rewards for intermediate andproactive strategies can be substantial, and incertain difficult areas this may be the only wayto achieve progress.

Furthermore, teams need to decide carefullywhether a more reactive or pro-active approachis most appropriate and feasible. A proactive ap-proach is not necessarily always superior. Infragile contexts or where political stability is atrisk, special care needs to be taken. A variety oflocal voices should be heard regarding whetheror not a more proactive strategy is advisable. Ata minimum, using GPE analysis to inform andshape strategies and operations is likely to leadto more appropriate Bank programs than hasbeen the case where programs have been in-formed exclusively by technical analysis andhave tended to disregard country and reformcontexts.



2. Note that local stakeholders are often not unified in their preferences (i.e. there is a greater variety to local preferences thanthe current official government declaration/policy). Various types of surveys can be used to tease out local preferences as wellas gauge how well preferences are informed by evidence/available information.

One can distinguish three broad levels of PGPEanalysis: the macro or country- level, the sectorand thematic level, and the project- and policy-specific level. GPE analysis can focus at one levelonly, or cover several levels, usually through aset of notes.1 This allows tailoring the analysis tothe specific needs of a team, as well as to the ex-isting context, including the fact that the den-sity of prior analytic work varies significantlyacross countries. Teams engaged in strategyplanning—for example, preparing a country as-sistance strategy (CAS) or a major DPL or re-engaging with a country—may want to coverthe country level as well as one or several keysectors. Task teams working on one particularoperation will be most interested in project- andpossibly sector-focused GPE analysis—whichideally can benefit from and draw on existingcountry-level analysis.

i. Country-level analysis seeks to capture theoverall governance situation and the main po-litical economy drivers. Analysis at this levelserves to capture important factors, such asthe geopolitical context, important social

divisions, including ethnic or religious oneswhere relevant, how these have evolvedover time (including the impact of pastlegacies on the current political and eco-nomic landscape), and the evolution of thepolitical management of economic rents.Such analysis is crucial for reaching a com-mon understanding of the country contextin which the Bank Group operates. Its mostdirect uses are to inform country assistancestrategies (or other strategies) and to serveas a background for a more specific GPEanalysis.2

ii. Sector and thematic PGPE analysis focuses onspecific areas and may cover an entire sector‘value chain’ (see Annex 2) or selected issueswithin a sector or (broad) theme. GPE analy-sis of sectors complements technical analy-sis, such as technical studies of infrastructureneeds or of capacity-building needs in thecivil service. The analysis can inform CASpillars, sector or thematic economic andsector work (such as CEMs for growth as atheme), as well as DPLs, investment lend-ing operations, or TA.


Part Three: Choosing the Level ofAnalysis: Country, Sector, Project

1. For example, most CGAC pilots combined a background paper on country-level issues with specific pieces on various sectorsor themes.

2. As set out above, most existing tools developed by other development agencies address this level.

iii. The third level is problem-driven GPE analy-sis focused on specific projects and/or policy de-cisions. Such analysis is focused directly ongenerating advice to a single operation orto aid in dialogue on a specific policy issue.

Importantly, the different levels can be andoften are combined. For example, country teamspreparing a CAS or major DPL may commis-sion an analysis that combines a summary ofcountry-level GPE dynamics as well as two orthree notes focused on specific sectors orthemes. Such a combination was used in theCGAC GPE pilots in Mongolia, Zambia, Mali,and Lebanon (see the right-hand information

in Figure 3). Similarly, sector task teams mayask for an analysis that sets out the sector-wideGPE context (and how it is linked to adjacentsectors and/or country dynamics), as well as ad-dressing one or two questions around specificpolicy reforms—such as those to be defined asPrior Actions or benchmarks for a DPL. Each ofthe three levels is described in greater detail inAnnex 2.

Ideally, PGPE analysis at any level shouldincorporate comparative thinking, even if themain interest is in one country or sector at atime. Comparative case studies are one of thefundamental methods used in the social sciencesto test causation.3 Comparisons help assure that

PA RT T H R E E


F i g u re 3 : Levels of analysis: country, sector, project

3. Originally based on the methods of inductive logic developed by John Stuart Mill. For recent discussions see King, Keohane,and Verba (1994); Ragin (1987).

Level 1

Level 3

Level 2

Macro-level countrygovernance and political

economy assessment

Thematic analysis Forexample: PGPE analysis ofNRM, of governance and

growth, etc.

PGPE analysis ofcore governanceissues and reforms

PGPE analysis ofother sectors (e.g.HD, transport)

PGPE to inform aspecific project orcomponent

Political economyanalysis focused on asingle policy decision

Thematic GPE assessments—e.g.of managing natural resources,linking governance and growthanalysis, political commitment topoverty reduction.

Sector GPE analysis inform CASpillars, sector strategies, sectorDPLs, complementing technicalsector analysis

Sets the stage, provides detailedqualitative view of governanceand political economy drivers forthe country overall. Often asbackground note to more specificGPE analysis.

Detailed, focused problem-drivenGPE analysis can inform specificoperations or components,specific policy issues (e.g.regarding Prior Actions).

Fo

cu

s o

f C

GA

C p

ilo

t G

PE

As

analysis does not reach erroneous conclusionsabout causality. For example, a sector analysismay conclude that the lack of strong demandfrom citizens is a major cause for the lack of re-forms in this area. However, in another sectorin the same country, reforms may have occurreddespite similarly weak demand. Consequently,the lack of citizen demand should perhaps notbe considered the central cause or driver in thefirst sector.

For Bank teams, real comparative studies areoften not a priority, given that they are inter-ested in one particular country or sector at atime. However, within-country comparisonsmay be possible and of interest, for examplewhen considering whether reforms have workedin other sectors in the same country, and whataccounts for such differences in reform success.Furthermore, it may be fruitful to explore

whether similar issues are being experienced inequivalent sectors in similar countries and/orwith similar types of project interventions.

For a number of issues, it may be worth ex-ploring whether countries with similar endow-ments and conditions have taken similar orrather diverging trajectories. This can help tocapture what factors are making a difference andwould merit attention through interventions.For example, two countries with comparablelevels of income may have started from similarlevels of literacy, but one country may clearly bemoving ahead of the other. Some comparativeexploration can help to tease out institutionaland sociopolitical factors that account for dif-ferent levels of commitment, effort, and resultsin expanding education, and this in turn canhelp to define what could be targeted in laggingcountries.

C h o o s i n g t h e l e v e l o f a n a l y s i s : c o u n t r y, s e c t o r, p r o j e c t


Credible GPE analysis needs to be well evi-denced. It needs to pull together a compelling‘analytic narrative’ consistent with experienceand systematically gathered data and informa-tion, while avoiding the pitfall of essentiallyproviding technical analysis combined with somebroad statements about governance and politi-cal economy. This balance needs to be struck ona case-by-case basis, as part of the process ofdefining the focus of the analysis, and based onthe needs and intentions of the country or taskteam. As discussed further in Part Five, problem-driven GPE work should whenever possible beclosely integrated with technically focused analy-sis and teams so that the two can be comple-mentary and so that important synergies foroperational implications can be reaped. Usefulways to structure the analysis and existing toolsand experiences can also be found in the 2007TIPS handbook (World Bank 2007c).

The three-layer model set out in Part Oneprovides a broad guide for the evidencingprocess. The first layer, the problem to be ana-lyzed, is usually what motivates the analysis,and hence would be reflected in the conceptnote or terms of reference (TOR). In manycases, it may be relevant to gather additional in-formation for describing the problem in detail,

and for getting a more precise grasp of the mech-anisms by which GPE factors contribute to thepoor outcome to be explained. For this, theGPE analysis is often able to draw on existingmaterials such as technical reports, investmentclimate assessments (ICAs), PERs, CEMs, orPAs, as well as on discussions with country, sec-tor, or task teams. At times the GPE analysismay bring out that further technical informationon vulnerabilities or challenges is needed.

At the heart of the analysis are institutionalstructures and governance arrangements, as wellas stakeholders and their interests and incentives,and how all of these interact.

i. To the extent that they are available, GPEanalysis should draw on any relevant pre-existing diagnostics, such as PETS, corrup-tion assessments, and Public Expenditureand Financial Accountability (PEFA) as-sessments.

ii. Important sources for capturing institu-tional structures are laws and regulations,as well as organigrams and mappings of dejure and actual process flows (the latter isessential to capture informal rules and ac-tual behaviors). For example, the analysismay map the formal as well as the actual


Part Four: Evidencing Problem-Driven GPE Analysis

processes for setting electricity tariffs, or forapproving public investment plans.

iii. The analysis may be able to utilize interest-ing sector-specific data (such as data aboutthe size and relative importance of SOEs,data on hospital usage, absenteeism, infor-mation on pricing, and so on), as well asinformation about government policies,implementation status, public sector ca-pacity in various areas, and so on.

iv. To the extent possible, the analysis shouldinclude coverage of informal institutions.This is essential for understanding whetherformal rules are actually functioning as in-tended or not—which in turn has impor-tant implications for the kinds of policyand institutional changes that Bank opera-tions and policy dialogue may seek.

v. Surveys of public opinion or of specificgroups (such as business people) or house-hold surveys that include governance-related questions may also be available andcan provide evidence about whether gover-nance arrangements are perceived to be de-livering or not, the prevalence of integrityproblems, and so on. Sometimes, commis-sioning a specific survey can also be useful,for example to understand what preferencesdifferent groups in society have regarding aspecific issue or policy.

vi. Other written sources that may be useful toanalyze include media coverage and recordsof parliamentary debates.

Apart from written sources, interviews withindividuals and focus groups (where appropri-ate) are a crucial source of information. It is im-portant to approach interviews systematicallyin order to extract key information. Interviewsand discussions with well-informed individualsare essential for developing an understanding ofpolitical economy drivers, especially for issues on

which little information is available either pub-licly or as the result of operational work. Themost valuable interlocutors are individuals whohave an overview of a situation and who are in-terested in discussing this information, for ex-ample because they share a conviction that re-forms are needed.

i. Interviews with knowledgeable local stake-holders—not only researchers, journalists,civil society representatives, but also policythinkers within government and politicalparties—are essential for building an un-derstanding of the issues and political econ-omy drivers. They should also yieldinformation about social norms and defacto (as compared to de jure) rules of thegame. Interviews can be done individuallyor as focus groups, with the former beingpreferable for more sensitive topics. Prepar-ing precise questions is essential to gather-ing a rich and pertinent set of informationthrough interviews.

ii. One-on-one discussions or workshops withBank country or sector team members andlocal staff are a good way to capture theiroften considerable (tacit) knowledge aboutpolitical economy issues.

iii. It is often good to snow-ball interviews:moving from a few initial interlocutors suc-cessively to a wider set. It is generally best tointerview key individuals once initial infor-mation has been gathered, to be sure thatthe most pertinent questions can be asked.

iv. For detailed stakeholder analysis, interviewsneed to focus on stakeholder interests, whatlevel of change they support or oppose,how much they care about an issue, andhow much leverage they have over otherstakeholders. Detailed information of thisnature is especially essential when formal-ized stakeholder analysis tools are being

PA RT F O U R


used (see Annex 1, Box 10), but it can bevaluable also for less formalized analysis.

A general core principle for evidencing GPEanalysis is triangulation. For interviews, thismeans posing a similar set of questions to mul-tiple respondents in order to corroborate claims.1

It is also helpful if more anecdotal or ‘soft’ infor-mation can be triangulated with ‘harder’sources—such as actual budget and public in-vestment allocations, information on differentialpricing, asset declarations of public officials, aprocess-tracing of a certain chain of events, ortracing published or publicly accessible sourcesthat reflect positions taken by individuals on theissue at hand (for example, on electricity tariffs,teacher training and wages, or governancearrangements for the financial sector).

It is important to recognize that detailed ev-idencing of problem-driven GPE diagnostics canbe challenging. Pertinent data or information isnot routinely collected by sources such as acountry’s national statistical office or ministry offinance, nor by World Bank teams/country of-fices. In addition, there may be issues with theaccessibility of relevant data. Evidencing can beparticularly challenging for sector-focused andthematic analysis. For country-level analysis, agreater amount of relevant information is oftenavailable through the national press or previouslypublished analysis. For project or policy specificPGPE analysis, evidencing is made easier by thenarrower focus of such work, which can oftenbenefit from easier access to specific data sources

and well informed people (especially if there isan existing operation and/or ongoing policy di-alogue).

Working with good local consultants is highlyvaluable to ensure well evidenced GPE analysis.Local consultants can be important to identifyand access useful sources of evidence, for ac-cessing such data and for providing a first analy-sis. They can also give advice on potential in-terviewees, on process issues, and on theinterpretation of the information gathered.Countries vary considerably with regard to howaccepted it is to discuss governance problemswith outsiders. Again, local consultants can bevery important and helpful for understandingand interpreting nuances. They may also play arole in the follow-up and dissemination (seePart Five for a discussion of these issues). Iden-tifying good local consultants for such workshould therefore receive high priority.

Inevitably, there is a trade-off between thedepth of evidencing and resources allocated. Anin-depth, richly evidenced PGPE analysis willrequire more time and resources than a quickpiece that is more conceptual, drawing on a fewkey interviews only. Teams commissioning prob-lem-driven GPE analysis should be realistic inwhat they expect relative to the time and re-sources allocated. On the other hand, teams un-dertaking the analysis will often need to be cre-ative in thinking about what could be usefulsources of evidence, seeking to be systematic incollecting information and bringing it to bear ontheir story line in a convincing way.

E v i d e n c i n g P r o b l e m - D r i v e n G P E A n a l y s i s


1. Attempts at triangulation can sometimes be challenging, since different interlocutors may contradict each other. In such cases,it is important to consider whether an interlocutor has an interest in raising a problem, exaggerating it, or denying/diminish-ing it.

Getting the process right is essential for PGPEanalysis to be useful and effective. Quality GPEanalysis benefits from a clear vision of the in-tention of the work. It should be clarifiedwhether it is meant to inform Bank teams in-ternally only, or whether it is also intended asan input for debates among country stakehold-ers. Likewise, the focus and the type of outputbeing sought should be clear. Also necessary is agood understanding of the process and issuesthat are likely to appear along the way. Further-more, it is vital to consider the follow-upprocess, how the analysis will feed into relevantprogramming, strategies, and operations, as de-scribed in Part Two above.

This section addresses the following fiveprocess issues1:

V.1 Planning problem-driven GPE workV.2 Defining and finding the necessary skillsV.3 Issues to consider regarding implementa-

tion of the diagnosticsV.4 Sharing and disseminating outputsV.5 Bridging analysis and follow-up/action.

This section is primarily focused on how toproduce a PGPE analysis that can effectively in-form Bank strategies and operations. In situa-tions where teams consider that PGPE analysisjointly with governments is feasible and desir-able, additional process considerations apply.Here, interaction with country stakeholders isprimarily raised with regard to follow-upprocesses, that is, the sharing and disseminationof PGPE diagnostics and the implications thatmay emerge for engaging differently with coun-try stakeholders.

V.1. PLANNING GPE WORK ANDLINKING IT TO OPERATIONS ORTECHNICAL ANALYSIS

As for other diagnostic work, good planning isimportant. Answering the question of why thiswork is being done and reaching clarity on theobjective of the analysis and the questions to beinvestigated are crucial.2 As the problem-driventhrust of this framework emphasizes, the focusof PGPE work can be grounded in experienc-


Part Five: Getting the ProcessRight: Issues to Consider

1. Further advice on these issues is available ‘on-demand’ from the teams located in PRMPS and the GAC secretariat.2. Agreeing on an inception report as a first output can be one way to achieve greater clarity as part of the process.

ing difficulties with prior operations, in seekingto respond to new development challenges(such as major shifts in property rights or deal-ing with regional inequality), or in diagnosingthe governance and political economy dimen-sions of other major questions addressedthrough ESW. It is helpful if the concept noteand TOR for the PGPE analysis clearly sets outthe scope and focus of the work to be under-taken For example, for sector work this can en-tail specifying which part of the sector valuechain (as discussed in Annex 2, section II.2.1.2)is to be analyzed.

As a general rule, problem-driven GPEanalysis will be particularly valuable if initiatedupstream of a strategy (CAS/ISN) or an opera-tion (DPL, investment lending, or major TA).Nevertheless, GPE analysis may also become amatter of urgency due to an economic crisis ora reform impasse or reversal, requiring real-time,rapid assessment, or when a window of oppor-tunity opens up, for example due to changes ingovernment, when Bank teams may be calledon to support major new reform drives.3 Fur-thermore, in the case of multiyear strategies oroperations, changes in events may trigger a needto update the analysis in order to provide aninput into how the strategy or operation maybest be adapted to the evolving situation. There-fore, at upstream stages, there is the greatestvalue in investing in substantial analysis. Duringimplementation, when the issue is whether andhow to adjust an operation or strategy, a brief notecombined with hands-on advice to teams/TTLs isoften most appropriate.

Country management’s commitment to andengagement with GPE analysis is important for

reaping its potential strategic benefits to countryprograms and sector operations. Country direc-tors or managers in many cases are the crucialhinge between undertaking the analysis and tak-ing decisions on the follow-up (as set out in PartTwo). Ideally, therefore, they should be involvedfrom the start.

Moreover, linking GPE analysis well to other(ongoing) analytic and advisory activities(AAA) and to Bank operations is likely to in-crease its ultimate usefulness. At a very practi-cal level this means involving sector/task teamsin the design of problem-driven GPE, estab-lishing links between the teams, and ensuringthat GPE teams are fully aware of existing andongoing technical analysis (including relevantdrafts). Often, CMUs can play an importantrole in this regard.4

PGPE analysis involves thinking through thekinds of data and information that will beneeded for meaningful work. As discussed inPart Four, PGPE analysis may use a wide rangeof information and data, some of which maynot routinely be collected by country offices, in-cluding opinion and other governance-relatedsurveys, a mapping of sector processes, pricinginformation, and so on. Some up-front work bythe local office or local consultants can be veryhelpful in ensuring that PGPE analysis is wellevidenced.

V.2. HOW TO DEFINE AND FIND THENECESSARY SKILLS

Getting the right skill set is essential to goodproblem-driven GPE analysis. The requiredskill set varies with the level and focus of analy-

PA RT F I V E


3. Of course, rapid assessments are easier to turn around quickly and effectively if there is prior work to build on. 4. It is especially important to pay attention to proper linkages and integration if PGPE studies are funded through TF rather

than being part of the overall AAA planning of country teams; as well as when PGPE analysis is being undertaken by exter-nal consultants.

sis. Especially for upstream substantial analyticwork, it is best to have a small team, bringingtogether a set of skills. Matrix 1 summarizes keyskills by level of analysis (the concrete skill re-quirements of each specific case may vary).

In most cases, meeting this skill set requires acombination of at least two people. The teammay be composed of one person with strong po-litical economy skills and another with detailedsector knowledge. Where part of the analysis iscarried out by consultants, it is important that

country or task team staff also be engaged (ashands-on TTLs), in order to ensure that the re-sulting work is relevant to the Bank as well asthe subsequent uptake of the analysis. Further-more, as a general rule, good local consultantsshould be included in PGPE teams wheneverpossible. Apart from their contribution to un-dertaking the actual analysis, local consultantscan help navigate the process of disseminatingthe analysis and the follow-up process where itconcerns local stakeholders.

G e t t i n g t h e P r o c e s s R i g h t : I s s u e s t o C o n s i d e r


Level of analysis Key skills

Country-level analysis Country and/or regional knowledge and the ability tosummarize key issues coherently and compellingly

A good grasp of key concepts—such as stakeholders,institutions, rents—and experience in applying them infocused analysis to developing-country dynamics

Thematic and sector-level analysis A good grasp of key concepts—such as stakeholders,institutions, rents—and experience in applying them toconcrete development challenges, such as sector reforms

Problem-solving, practical orientation; including anunderstanding of Bank operations and an appreciation ofdevelopment effectiveness concerns

Experience in analyzing public sector issues, institutions,and processes (e.g., public investment, public sectoremployment issues) with cross-sectoral relevance

Knowledge about typical governance problems related tothe theme or (sub-)sector to be analyzed and how theyare linked to political economy drivers

(Team lead) An ability to structure a task in terms of theanalytic/investigative process and to produce a well-structured output

Project-focused analysis Problem solving, practical orientation, includingfamiliarity with Bank operations

Ability to apply key GPE concepts to project-level issuesand challenges

Depending on the nature of the project and issue:familiarity with focus groups/participatory approaches

M a t r i x 1 : Levels of analysis and key skills needed

Experience with recent studies suggests thatPGPE analysis benefits from a well-placed TTL.The TTL should be in a position to link the di-agnostic work to the operational challenge it ismeant to inform (a CAS, DPL, TA, etc.). An-other important consideration is whether theTTL can commit sufficient time to ensure goodplanning, quality management, finalization,and dissemination. Ideally, the person leadingthe analysis would also be engaged in the actualfollow-up of shaping operations and strategies,although this may not always be feasible. Ulti-mately, what works has to be decided on a case-by-case basis.

Once the required skill set has been defined,finding the right team is the next challenge.Problem-driven GPE-type analysis is still a rel-atively new area, so the number of staff and con-sultants with the desirable combination of skillsand experience is still limited. However, thepool of people with the requisite experience isexpanding continuously. The PREM anchor, re-gional PREM units, and other network anchorsare often able to provide advice and recom-mendations. Some broad categories of potentialconsultants for problem-driven political econ-omy work are summarized in Matrix 2.

PRMPS is currently working on building aroster of relevant consultants and the specificskills which they can bring to a team (with afocus on core political economy and core gover-nance skills). This roster will complement rostersbeing created by Bank regional and sector teamswith a governance and political economy focus.

V.3 IMPLEMENTING GPEDIAGNOSTIC WORK AND ENSURINGQUALITY MANAGEMENT

Many of the challenges associated with imple-menting problem-driven GPE diagnostic work

are similar to those for other AAA (such as in-suring timeliness and collaboration with localconsultants/researchers to foster local capacity).However, there are at least four specific chal-lenges worth considering:

• First, team involvement and openness are cru-cial. The PGPE analysis will be more truly‘problem driven’ and more focused if teamsare willing to discuss what challenges havearisen with relevant country or sector pro-grams in the past. Also, it is important thatteams and local staff take the time to discusstheir own perceptions of relevant politicaleconomy issues—of which they often haveconsiderable tacit knowledge.

• Second, PGPE analysis requires a decisionwhether to inform government explicitly of thework being undertaken. As a general rule,teams have found it helpful to be open and toexplain the motivation of such work, but theultimate decision is on a case-by-case basis.

• Third, it is essential to ensure good feedback tocountry/sector/task teams. This is essential tovalidate the findings and to engage in theprocess of translating findings into opera-tional implications and follow-up (see alsoPart Two).

• Fourth, it is important to consider (especially)the downstream quality management for GPEanalysis. The basic components of qualitymanagement are upstream concept notepreparation and reviews of draft outputs.Planning the downstream quality manage-ment process is essential for finalizing theoutputs and for ensuring that there is aprocess for review and approval for such a(relatively) new type of work. It is also im-portant to facilitate the production of out-puts that can be shared within teams or morewidely.

PA RT F I V E




Type of consultant/affiliation Typical strengths/weaknesses*

University-based political scientists Strengths: can bring excellent country knowledge, orstrong conceptual skills, and in some cases both; usuallygood writing skills.

Weaknesses: experience with consultancy work varies; and(detailed) knowledge of development policy issues andoperational approaches is often limited.

Researchers/consultants based atthink tanks and consultants basedat firms

Strengths: increasing pool of specific experience withsimilar type of work also for other agencies; goodknowledge of development issues and operationalapproaches/aid issues in general; quality managementwithin the firm/institute can reduce the need to micro-manage the product (once a contract is established);professional approach to handling the task and producingan output.

Weaknesses: specific country knowledge may be morelimited than among specialized academics; firm-basedconsultants, especially, have some incentive to applycookie-cutter approaches; tend to be more costly thanindividual STCs.

Individual STCs with differentbackgrounds (including Bankretirees)

Strengths: vary; often know very well how the Bankoperates (especially retirees and long-term STCs).

Weaknesses: vary.

Regional consultants with varyingbackgrounds (e.g., in auditing andaccounting vs. think thanks)

Strengths: can have a greater intuitive understanding ofissues; familiarity with typical problems in the region, e.g.in sectors; language skills.

Weaknesses: report structuring/writing skills vary; possiblysome reluctance to tackle difficult political economyquestions.

Local consultants (with variousaffiliations)

Strengths: detailed country knowledge (althoughattention needs to be paid to familiarity with the actualissue at hand); can have good contacts in country andfamiliarity with how to access various data andinformation

Weaknesses: report structuring/writing skills vary; risksrelated to interpreting TOR can be greater—so discussingTOR is important; sometimes serious problems with over-commitment of good local consultants resulting in delaysin delivery; may have their own political biases whichneed to be taken into account.

* Note: these are very much generalized and may differ greatly depending on the individual consultant.

M a t r i x 2 : Types of consultants for problem-driven GPE analysis

V.4 SHARING/DISSEMINATINGOUTPUTSThe question whether and how to share and dis-seminate PGPE analysis is often a crucial andsensitive one. At one extreme, PGPE work is un-dertaken, but its outputs are only shared withtwo or three key members of the country team,because of the perceived sensitive nature of suchreports. Other work has been published asworking papers or as chapters of ESW or otherformats (see for example: Campos and Syquia2006; Lal 2006; World Bank 2008a). DfID hasmade many of its Drivers of Change studies pub-licly available through an outsourced knowledgerepository website that it sponsors.5

As set out below, there is a wide range of op-tions between keeping circulation limited to afew people only and fully publishing a docu-ment. Very limited circulation has a consider-able cost in terms of reducing the usefulness andeffectiveness of any PGPE work, and should

generally be considered only if the document isextremely sensitive for some reason.

PRMPS is currently developing an internalSharePoint site, which will allow invited WorldBank users to access PGPE-related materials andreports.6 The intention is to provide a spacewhere the confidentiality of such documentscan be protected, but that at the same time al-lows an effective community of practice toemerge across the institution.

Following are further considerations and op-tions for sharing/dissemination:

• Country context and the relationship betweenthe Bank and country stakeholders vary greatly,and dissemination choices are dependent onthese specific contexts. The default optionshould be to seek wide(r) dissemination, butthis may not always be possible (see alsoDAC principles on Donor Approaches toGovernance Assessments 2009).7

PA RT F I V E


B o x 6 : Dissemination experiences of pilot teams

In most countries where PGPE-type analysis has been undertaken in recent months, teams haveeither disseminated synthesis notes to government and other stakeholders, or are planning to doso; and they also have informed governments of the intent to undertake such analysis up front.In contrast to individual earlier experiences, no country team involved in these recent processes hasexperienced major difficulties as the result of undertaking GPE analysis; and the goals of improvingdevelopment effectiveness and understanding country context better are appreciated by mostlocal stakeholders.

Nonetheless, many CGAC/GAC for CAS pilots have grappled with challenges in dissemination.The first steps have been intensive interaction and feedback to country and task teams throughone-on-one discussions, circulation of the diagnostic notes to task team leaders, and workshops.In Uzbekistan, key points of the governance diagnostic were disseminated to a wide range ofstakeholders as part of the CAS consultations. Outputs from the Indonesia, Mongolia, and Zambiaanalytic work have begun to be disseminated especially through PowerPoint presentations to avariety of internal Bank audiences, as well as to other development partners.

Source: Authors.

5. See: http://www.gsdrc.org/go/topic-guides/drivers-of-change#cam. 6. See: http://connect.worldbank.org/units/prem/ PD-GPEA/pdgpe/default.aspx (accessible via World Bank intranet only). 7. Principle 5: “Making results public unless there are compelling reasons not to do so.” DAC/GOVNET (2008) Donor Ap-

proaches to Governance Assessments, Guiding Principles for Enhanced Impact, Usage, and Harmonization, Final Draft.

• A central aim should be to ensure that GPE di-agnostics reach their primary audience—oper-ations staff and managers in country, sector, ortask teams responsible for shaping relevantstrategies or operations. As discussed above,ideally members of relevant staff teams willhave been involved in generating the GPEanalysis, but it is important to disseminatethe results among teams as well as manage-ment in order to promote reflection on im-plications and discussions about follow-up.Furthermore, in-depth country or sector levelanalysis should not only reach current staffbut also be helpful for transferring GPE in-sights when staff changes— which frequentlyhappens while earlier designed strategies oroperations are still under implementation.

• Generally, it is desirable to share at least a syn-thesis output (that is not overly bland) with arange of stakeholders, including the countrygovernment and nongovernment stakehold-ers. This helps to reap the full benefits fromundertaking the analysis, and helps preventany misleading assumptions. Shareable ver-sions are also useful for interacting withother development partners in-country.8 In-country workshops can offer an opportunityto engage with a range of stakeholders in dis-cussing and validating findings and in draw-ing out implications.

• More sensitive information can always be madeavailable as a background note or exclusivelyduring workshops. In general, the role of GPEanalysis is to analyze how governance arrange-ments and political economy drivers affect Bankoperations, rather than arguing ad personam orproviding any kind of ‘forensic’ diagnosis.

• Certain types of data are by nature less contro-versial. Public opinion surveys on governancearrangements, diagnosis of certain policy is-

sues, or the effectiveness of accountability—these can be shared as a more ‘neutral’ formof evidence as well as a way to corroboratejudgments.

V.5. MOVING FROM ANALYSIS TO FOLLOW-UP

Creating an effective link between analysis andfollow-up requires attention and the timely ini-tiation of a subsequent process. As set out in PartTwo, there are a number of ways in which prob-lem-driven GPE analysis can be used to informand shape strategies and operations. The processof moving from analysis to follow-up usuallystarts once initial complete drafts of the analysisare available; it should not wait until the writtenreports are fully finalized. A workshop with thecountry/sector/task team will often be useful atthis stage to review and debate the findings andto discuss what implications these findings havefor strategies or operations.

There is therefore a dual review process forPGPE analysis: one review focused on the qual-ity of the product itself and the other focused onwhether and how to take the analysis on board.Especially if GPE analysis is used to inform asector where reforms have been stalled or re-versed, defining implications may require an in-tense, detailed, and in-depth discussion betweenthe GPE analysis and the sector team.

A discussion about the implications and fol-low-up can focus on the range of options be-tween a more reactive and a more proactiveapproach, as described in Part Two. In particular,if there is an intention to pursue an intermediateor more proactive approach, there is then a needto reflect on how this can be realized. To date,such follow-up work has often drawn on trustfunds and other kinds of non-core resources.



8. See also an interesting summary of the experience with a multi-donor governance assessment based on the EU’s governanceprofile (i.e. country-level diagnostic) in Cameroon http://www.oecd.org/ dataoecd/24/27/40094521.pdf.

This good practice framework has sought to setout some key points regarding concepts and ap-proaches for problem-driven governance andPGPE diagnostics, and some issues and experi-ences regarding how to implement and use suchdiagnostics.

Three points are especially worth reiteratingand emphasizing:

• The ultimate objective and motivation forsuch analytic work is to develop more feasi-ble approaches to reforms in order to en-hance development effectiveness.

• The work can be most successful if teams areopen to and actively engage in this diagnos-tic work, articulating needs and a clear focusand pursuing the synergies between technicaland GPE-type diagnostic work, as well as en-gaging in frank discussions of tacit knowledge,past experience, and upcoming challenges.

• There is an emerging community of practiceboth inside and outside the Bank, whichteams pursuing such work can tap into (andbecome part of ).

Given the rapidly expanding interest in suchwork and the scaling up that has been enabled,not least by the Governance Partnership Facility,it is expected that a much richer set of experi-ences will become available in the course ofFY2009–10. The intention is to update thisframework based in due course as new experi-ence accumulates and new needs emerge.

Feedback from task teams regarding thisframework is therefore strongly encouraged andmay periodically be actively sought by theGAC/Public Sector Governance team.


Part Six: Conclusion and Looking Forward

A wide-ranging and fast-growing academic so-cial science and policy literature provides in-sights and applied investigative methodologiesinto the political economy of developing coun-tries.1 This annex highlights some essential con-cepts and approaches relevant for PGPE analy-sis.2 The annex is not meant as a comprehensiveliterature review; rather, it flags key issues andapproaches to operationalizing political econ-omy analysis to help support engagement by in-ternational development partners such as theWorld Bank. The aim is to provide a succinctoverview of key approaches and concepts, whichin turn can be complemented over time with rel-evant case practice from applied analysis in thisarea. The approaches discussed here are partic-ularly relevant for targeting the second and thirddiagnostic layers: governance arrangements andunderlying political economy drivers.

I.1. THREE CLUSTERS OF DRIVERS:STRUCTURES, INSTITUTIONS, ACTORS

Three key types of factors or variable are com-monly considered in PGPE analysis: structures,institutions, and actors (see Figure 4). Thesefactors are relevant for the second and the thirdlayers of the analysis. Analyzing institutionalvariables most directly informs the second layer(governance arrangements), while the dynamicsbetween the three types of variables and the in-centives and interests they generate for ac-tors/stakeholders are crucial for analyzing un-derlying political economy drivers.

Structural factors can be understood asdeeper features that affect the political econ-omy of the respective country. These tend tochange only slowly over time and are beyond thedirect control of (local) stakeholders. One such


Annex 1: Political Economy: Key Concepts and Approaches

1. Selected background readings include Weingast and Wittman (2006), Rhodes et al. (2006) and Goodin and Tilly (2006). 2. The section covers more approaches developed in political science than in economics. Many political economy models that

have recently emerged in economics are more focused on (highly) developed countries—which are marked by strong formalinstitutions/clear rules of the game, highly developed markets, and established democratic and electoral systems. There are in-creasingly overlaps—driven inter alia by economists focused on developing countries and development issues such as Dani Ro-drik and Mushtaq Khan.

feature is geographic and resource endowments.For example, high dependence on natural re-source exports can subject countries to thevolatility associated with commodity prices, andas highlighted in the literature, the nature of re-source endowments can significantly shape the

development and institutional trajectory ofcountries. The analysis may therefore wish to re-fer to relevant comparative literature discussingthe potential relevance of particular structuralfactors for shaping and interacting with pre-vailing institutional/governance arrangements.3

A N N E X 1


F i g u re 4 : Three clusters of drivers

Structural variables

Examples:Economic base and level of development

Climate and geography (including effects of climate change), and geopolitical situation

Nature of interaction with the global economy (trade, migration, etc.)

Population dynamics; urban-rural relations; types of urban centers

Status of poverty and of equity/inequality

Institutional variables

Examples:Macro: constitutional set-up, electoral rules, major laws

Detailed: rules governing policy and budget processes, ‘organizational’ institutions: set-up of government; ministries and their roles and mandates; accountability institutions (parliaments, SAIs, etc.) and interaction between these

Informal: social norms and expectations; nature and strength of patronage networks

Influence political and public sector action andpolicies and their implementation

Outcomes

Growth, poverty reduction, human development,dealing with development challenges

(pollution, social conflict, etc.)

Actors/Stakeholders

Examples:Detailed: political leaders/leaders in a bureaucracy/heads of SOEs; mid-level bureaucrats;

Macro: political parties, (organized) interest groups—business associations, trade unions, religious groups, farmers associations, CSOs, etc.

External stakeholders—other governments, international networks, development partners, etc.

3. For example, the recent ‘resource curse’ literature has highlighted the impact of the specific nature of a resource endowment(e.g., point versus diffuse nature resource). The specific type of diamonds endowment in Botswana has been associated withthat country’s relative success.

Other structural factors may be a high level ofinternational migration and reliance on remit-tances (see the case of Moldova in Box 7).

Institutional variables are those related to‘the rules of the game’ (laws and regulations, aswell as informal rules, such as social obliga-tions). Institutional variables may be of overar-ching importance in the political economy of asetting. In a wider sense, these variables alsocomprise the ways in which a public sector is or-ganized. Depending on the focus and level ofanalysis (country or sector or operation), the in-vestigation may focus on institutions of rele-vance to a particular ‘problem-area’—such asthe regulatory arrangements in the electricitysector. At the country level, the constitutionalset-up and regime type are key institutional vari-ables. Also, electoral rules can be of considerableinterest (see Annex 2 below), as they create in-centives for the kinds of policies and budgetaryresource allocations that politicians seek. Simi-larly, rules governing the policy process (or their

weakness/absence) can be highly relevant forthe degree to which private interests can influ-ence the content of laws.

Actors or stakeholders comprise individuals aswell as organized groups such as political parties,the military (and, in some cases, powerful secretservices), business associations, NGOs, tradi-tional associations, etc. Importantly, externalstakeholders, such as donors, foreign policy ac-tors, or foreign investors, often also play im-portant roles, for example, in negotiations overpolicies, rules for foreign investment and foreigntrade, or in cross-border water and transportmanagement.

As Figure 4 reflects, the three types of vari-ables can interact in numerous ways. For exam-ple, shifts in international commodity pricescan have major repercussions for the nature ofrents and rent distribution among actors in agiven country. A key tenet of institutional eco-nomics is that (formal) institutions constrain orchannel the ways in which stakeholders behave.

Po l i t i c a l E c o n o m y : Ke y C o n c e p t s a n d A p p r o a c h e s


B o x 7 : Moldova Governance Assessment—migration and in-country stakeholder and institutional dynamics

The Moldova GAC assessment looks at the interaction of structural factors, institutions, andstakeholders with a view to informing the new Country Partnership Strategy (CPS). Moldova isamong the poorest post-communist transition countries, but growth has been relatively strong inrecent years.

In terms of structural factors, the analysis emphasizes the ambivalent role played by labormigration. On the one hand, migration and remittances have contributed to growth as well as topoverty reduction. On the other hand, the fact that very large numbers of dynamic, working-agecitizens have left the country combined with the fact that they provide an informal social safetynet for those staying behind significantly reduces the demand on the government to perform andto be accountable.

This has facilitated a situation in which key actors in the executive shape core institutions inways, which promotes their hold on power, while risking making these institutions less effectivein delivering public services. The whole situation could come under pressure to change if a globaldownturn reduces opportunities for migration and/or opportunities to allow informal safety netsthrough remittances.

Similarly, a global downturn and decline in commodity prices could have major impacts oninstitutional and stakeholder dynamics in other CGAC pilots such as Mongolia and Zambia, whichare both major exporters of copper and other minerals.

Source: Jana Kunicova, ECA; authors

However, because formal institutions tend to beweaker in developing than in developed coun-tries, stakeholders more frequently seek tochange formal rules (trying to reform them, ortrying to shape them more to their own advan-tage); and it is also more frequent that formalrules become largely unobserved (such as the re-quirement to produce annual audit reports),and hence the exercise of private interests bythose in power is less constrained.

The framework in Figure 4 presents a stylizedway of thinking about governance and politicaleconomy and the interaction between the threesets of variables. Thinking in terms of variablesand how they interrelate can be helpful for struc-turing information and data, and for distilling a‘story line’. Such a framework can be used forframing the concrete, problem-focused analysisand for structuring the inquiry process (while itmay not always be suitable as an outline of theresulting report or output).

A crucial task of PGPE analysis is to capturehow the interactions between these variablesimpact on the problem or vulnerability at hand,including dynamics and trends. Box 7 gives abrief summary of how the interactions betweenthese variables were captured in the MoldovaCGAC analysis.

I.2. CAPTURING HISTORICALLEGACIES

Historical legacies often have profound effectson shaping current dynamics in countries wheredevelopment partners are engaging. Capturingsuch longer processes, and how societies con-tinue to deal with them, provides depth and per-spective to the issue of ‘how things have becomethe way they are today’. Such analysis capturingthe longer run will typically mean reaching backseveral decades, and possibly even to pre-colonialperiods where applicable. The analytical chal-

lenge of this type of work lies in summarizingthe key lines of evolution, and the interplay ofstructures, institutions, and actors (includingexternal ones) in shaping these lines. It meansasking for a summary of key trends, events,processes, and policies which give shape to thecurrent situation—while it is not a detailed his-torical analysis that would trace the many twistand turns of a country’s history over the pastcouple of decades.

The relevance of historical legacies for currentdevelopment challenges may often not be suffi-ciently explored. World Bank teams frequentlywork in countries which have received limitedattention from academic scholars, or where ex-isting attention has never really explored thelink between legacies and present constellations.Especially in settings where analysis of countrycontext is missing or scarce, it can be valuable toinclude such a perspective into GPE analysis aspart of understanding the broader context ofBank strategies, operations, and engagementproblems.

In particular, past conflicts and tensions maybe important to understand, as they can shapeattitudes long beyond their original occurrence.Examples are the legacies of war and tensions inthe Congo, the Balkans, or the Middle East. In-ternal conflicts between minorities and majori-ties can also cast long shadows. In many coun-tries, colonial histories and the decolonizationexperience have shaped institutions as well as ex-ternal relations in ways which still exert an in-fluence on political stability or a state’s institu-tional features today. In formerly communistcountries, it is often important to have a graspof how privatization proceeded—which formost countries involved massive re-distributionsof wealth, as well as sharp departures in terms ofsocial security versus opportunities—and howthis was linked to the process of political trans-formations. More generally, historical patterns of

A N N E X 1


how economic and political power have been al-located are important, as are the ways in whichpast equilibria were challenged or changed.

I.3. INSTITUTIONAL ANDSTAKEHOLDER MAPPING ANDANALYZING INFORMALINSTITUTIONS

In the short to medium term, the main dynam-ics are those between institutions and stake-holders. As set out above, the two are linked asactors shape institutions (especially where formal

institutions are weak/volatile) and, vice versa, in-stitutions influence the incentives and con-straints that stakeholders face.

I.3.1. Mapping institutions

Mapping institutions is valuable in three ways:firstly, it creates a clearer understanding aboutthe setting in which stakeholders operate, andhow this shapes their interests and incentives.Secondly, especially in areas of new or expand-ing operations in a sector, institutional map-ping enables a clear understanding of the coun-



B o x 8 : Formal and informal institutions

“Informal institutions—family and kinship structures, traditions, and social norms—not only matterfor development, but they are often decisive factors in shaping policy outcomes in environmentsof weak states and poor governance structures.” (OECD, 2007)

Formal institutions are codified—e.g. as laws and regulations—and usually have formalsanctioning mechanisms to make them effective. Informal institutions are norms and socialpractices rooted in history and culture; but they can also emerge as the result of the weakness,erosion or collapse of formal institutions. The enforcement mechanisms for informal institutionsrange from an adherence to internalized norms and expectations of reciprocity, to social shunningand ostracism, to threats and the use of violence.

Informal institutions exist in any society, as not all rules can or should be codified. However, therelationship between formal and informal institutions is very important, because their interactionoften shapes outcomes such as policy decisions or the implementation of policies. There are fourrelationships:

i. Complementary—i.e. informal institutions support the effectiveness of formal institutionsii. Accommodating—there is an acceptance of formal institutions, but informal institutions

circumvent these to some degree. iii. Substituting—informal institutions fill a void that is left by missing or incomplete formal

institutions. Informal credit markets and informal ways of contract enforcement are examples. iv. Competing or subverting—informal institutions diverge from, contradict, or contravene formal

institutions. Budget adjustments based on phone calls from the president or other powerfulpoliticians that contravene formal budgeting rules are one of many examples.

A general benefit of institutions for growth and development is that they lower transactioncosts, ensure that collective action dilemmas are solved and public goods are supplied. This worksbest when formal and informal institutions are complementary and worst when formal andinformal institutions are competing and mutually subversive.

Inefficient institutions (and interactions of formal and informal institutions) can persist becausethey are linked to power structures and to distributional benefits. For example, preserving (wide)informal powers can yield high immediate benefits to a country’s president. Also, a poorinstitutional environment can create economic profits for some activities that would be lost ifinstitutions are changed.

Sources and further reading: OECD (2007), chapter 2; Helmke and Levitsky (2004), Rodrik (2008b), Ostrom (2005),Fukuyama (2007).

terpart (e.g. a ministry of agriculture, or the na-tional and sub-national health sector bureau-cracy, or a national audit office). Thirdly, insti-tutional mapping is valuable to identifypotential levers of change. Institutional mappinginvolves a look at rules (especially formal rulesas a first cut) and at organizations relevant to aparticular issue, such as a sector, a policy and/ora broader development challenge (e.g. povertyreduction, dealing with environmental degra-dation, etc.). For institutional and stakeholdermapping, it is important to consider not onlythe relevant organizations and stakeholders (suchas ministries, agencies, individual leaders, lobbygroups, CSOs, etc.), but also the rules that gov-ern their interactions.

Such rules include formal laws and regula-tions, established processes, and the like; as wellas informal rules that are more derived from so-

cial norms and practices (see Box 8).4 In manyclient countries, formal laws and regulationsplay a very ambiguous role. On the one hand,laws can be significant, and may be frequentlyinvoked. On the other hand, not all laws are en-forced equally, or a particular law may be en-forced partially or selectively only (e.g. tax laws).Frequently, the implementing regulations for alaw to be actually effective are incomplete, or ex-isting regulations may be little known—evenwithin government agencies; or they are difficultto access, even if they are very important for cit-izens (such as regulations on how to registerproperty).

Sector, policy, and project focused PGPEanalysis can help to systematically analyze theweb of rules and how they are actually appliedin the current situation. It is important to un-derstand the current governance arrangements

A N N E X 1


B o x 9 : Analyzing winners and losers of electricity sector reforms in India

Power sector reforms have been on the agenda for a long time in India; yet they have sufferedfrom persistent gaps between political rhetoric in favor of reform and sector expansion, actualpolicies adopted, and implementation. The 2006 study by Sumir Lal seeks to understand thissituation from a political economy perspective, paying particular attention to the issue of ‘winners’and ‘losers’ from such reforms. Reducing subsidies to farmers for using energy to pump water isone of the key sticking point of these reforms.

As the analysis points out, the (numerous) poor as well as the (fewer) wealthy farmers could bepotential winners of reforms. The poor would gain if access to electricity would be expanded (ascurrently few of them have access), while wealthier farmers could benefit from more reliablepower supply. However, as the study points out, all farmers face a great deal of uncertainty: thepoor, that higher tariffs would actually result in the reallocation of funds to allow better access,or to fund other services better, and the wealthy farmers are uncertain about whether morereliable supply would materialize. Furthermore, farmers face uncertainty across multiple issuesand policy areas, including other inputs such as fertilizer and seeds.

Finally, the study finds that reforms in India have worked best if they have been gradual, andhave involved some ambiguity, which allows politicians to devise early gains for some, useopportunities arising out of the fluidity of interest groups, engage in compromises whileincrementally shifting the agenda, and so on. Thus, the pace and path of reforms had to respondto the more amorphous and complex situation, rather than being able to interact with clear-cutwinners/reform supporters and losers/reform opponents.

Source: Lal (2006).

4. The former are often referred to as the de jure rules; while the combination of formal and informal rules governs de facto theway in which things are done.

and what drives them, prior to mapping outwhat ‘better practice’ a country should adopt,and prescribing a potential change process.

I.3.2. Stakeholder analysis

Stakeholders come in many guises—stakehold-ers can be individuals or they can be specificgroups (such as mid-level officials of a ministry,or business or farmers’ associations, politicalparties, etc.). Three main ways of categorizingstakeholders have become widely used—‘de-mand-side’ versus ‘supply-side’ actors/stakehold-ers, reform champions versus reform opponents,and ‘winners’, ‘neutrals/undecided’ and ‘losers’.Such categorizations can be useful as long as theydo not lose sight of ambiguity and dynamics.

For example, MPs can be categorized as ‘de-mand side’ actors in some situations, but theycan also be part of the supply side of policies andthe supply of (good or bad) governance in gen-eral. Similarly, the categorization of certainstakeholders as ‘reform champions’ or ‘reformopponents’ often needs to be nuanced. For ex-ample, some stakeholders may favor a change ofthe status quo—but may nonetheless seek pol-icy reforms that diverge from the preferred pol-icy options proposed by WB task teams.

Furthermore, ‘reform champions’ often havemultiple agendas and issues which they need totake into account—maintaining their own po-sition and/or power bases, possibly addressing arange of development challenges (which may re-quire horse-trading and compromising with oth-ers), having important personal interests or ob-ligations, and so on. Also, some reformchampions may be strong individual supportersof reform, but are poor at coalition-building(e.g. politicians heading small pro-reform par-ties) which can limit their effectiveness. Under-

standing reform champions, their needs, incen-tives, and constraints can be very important forestablishing a relationship that really can bearfruit over time.

A specific sub-type of stakeholder analysis isthe ‘Expected Utility Stakeholder Analysis’,which uses a special software program ‘Sentu-rion’ to analyze data about stakeholders andtheir positions. Several World Bank teams haveused it on a pilot basis in recent years (see Box10). This is an interesting tool that can be pow-erful in certain situations—especially in order todrill down into how stakeholders will relate to avery specific policy issue or proposal, or wherecompromise may be found in the negotiationsabout a reform bill. Generally, the tool builds onpre-existing systematic qualitative analysis. Ifextensive information about stakeholders andtheir positions on an issue is available, it can bedeployed for drilling down into specific ques-tions and for illustrating likely processes andpoints of agreement relatively rapidly. A key is-sue to consider is the relatively high cost of us-ing this tool due to the program license and spe-cific training required for using the tooleffectively. Especially for countries with largerprograms it can be a worthwhile investment ifthe tool is then used to inform teams on a num-ber of issues.

I.4. RESOURCES, RENTDISTRIBUTION, POLITICALSTABILITY AND LEGITIMACY

Understanding the political economy of rents,and how it affects growth, poverty reduction andreform processes in general, represents a criticalpart of problem-driven analysis. Economic rentsemerge in a number of ways, and in turn sig-nificantly shape the incentive for actors.5 The



5. The standard definition of economic rents refers to the difference between what a factor of production is paid and how muchit would need to be paid to remain in its current use.

sources and allocation of rents and how theseevolve and shift over time are crucial for under-standing underlying political economy drivers,especially at the country level, and in particularways at the sector and policy level. Economicrents can be derived from two main sources:

those linked to control over natural resourcesand those derived from government regulation.The latter type of rents are those that AnneKrueger has drawn attention to in her famous1974 article, “The Political Economy of theRent-Seeking Society.” The latter type of rents

A N N E X 1


B o x 1 0 : Expected utility stakeholder analysis

Using a thorough data collection and recording process, the model translates the qualitativepolitical economy knowledge of country and technical experts into a dynamic quantitative analysis.This approach can be very helpful in deepening PGPE analysis around specific issues in a rigorousway. Using sophisticated modeling techniques, the approach simulates the complex bargainingdynamics among a large number of stakeholders. Visual aids embedded in the modeling softwaretransform raw data into easily understood patterns that are not readily apparent from the dataalone; and can help to visualize whether reform coalitions are likely to exist or emerge, and whatchanges in strategy a task or country team might adopt in promoting the emergence of a coalitionaround a viable degree of reform.

The agent-based stakeholder model has been developed into a user-friendly IT application,known as Senturion.

Agent-based stakeholder analysis is carried out in five steps:

1. Framing the Question: The framing step involves articulating the overall objective of thepolitical analysis problem and breaking it down into its component parts for detailed analysis.

2. Specifying the Issue: An issue, for the purposes of Senturion, is the continuum of optionsthat stakeholders may support on a specific policy matter. This is specified in a linear mannerfrom 0 to 100, either from the least politically difficult reform (0) to the most politically difficult(100), or in terms of two alternative poles of reform, with intermediate reform steps identifiedalong the spectrum in either case. Issue specification is probably the most challenging aspectof applying the agent-based model, and will often draw on prior qualitative analysis. Takingtime to specify the issue properly is also one of the model’s benefits, as it helps the teamachieve clarity on the reform issue, desired outcomes, and key stakeholders.

3. Collecting Data: The data collection process involves interviewing individuals with a stronggrasp of the country context, the stakeholder landscape, and/or the policy issue beinganalyzed. Interviews can be highly structured or broadly contextual, depending on the expertand the situation.

4. Analyzing the Basecase: The data described above are entered into the model, which thengenerates—using an algorithm that simulates bargaining dynamics—a picture of expectedstakeholder interactions over time. This is the model’s “basecase” outcome, resulting in agraphical display of initial results which can be used to analyze whether stakeholders areexpected to converge in support of a particular reform position, to what degree, and howquickly; and therefore what level of reform could be expected under the current dialogue andwith what level of support.

5. Defining and Interpreting Alternative Scenarios: Based on the above analysis, the teamruns simulations of strategies the Bank might be able to use to improve reform outcomes fromits own point of view. This analytical exercise allows the team to test whether an altered Bankposition could help achieve greater consensus in support of reform, or if the Bank couldrealistically exert leverage on a particular stakeholder to garner support for more extensive policychange. The model can thus provide a detailed illustration of the micro-politics surrounding theimplementation of reforms, and the costs and rewards of alternative Bank strategies.

Source: Naazneen Barma, Public Sector Specialist, East Asia Pacific.

can be adjusted through changing regulation,and in principle can be minimized—althoughthere may be arguments for the deliberate cre-ation of certain regulation-based rents, e.g. toprovide incentives for innovation.6 A furtherform of rents are those associated with illicitactivities—such as narcotics production andsmuggling or human trafficking, which govern-ments seek to limit.

Rents related to natural resources—such asminerals and oil, but also land and water—canbe very important, and in contrast to purelyregulatory rents, they cannot be eliminated. Fur-thermore, the size of these rents often lies to aconsiderable extent outside the control of localstakeholders, as with regard to fluctuating com-modity prices. Good regulation is important togovern the allocation and management of theserents, but achieving this can be extremely chal-lenging. In principle, globalization and increas-ing global populations mean that natural re-sources become increasingly scarce and valuable.Both growing scarcity and volatility of the valueof natural resource endowments and associatedrents can have a profound impact on the polit-ical economy and power relations within andacross countries.7

There are two competing arguments regard-ing rents: a more normative ‘good governance’perspective generally advocates that regulatoryrents should be reduced (or used for specificpurposes only), while rents resulting from nat-ural resources should be regulated in a waywhich maximizes (long-term) growth andpoverty reduction.8 A different argument—

made by Khan and others (e.g., Haber et al.2003)—holds that the distribution of rents is animportant pillar of political stability. External at-tempts to curtail or redistribute rents, or inter-nal contest over the distribution of rents can re-sult in political instability, and in the worst casecan lead to violent conflict. By implication, at-tempts to change rents should be approachedwith caution especially in potentially or actuallyfragile states. On the flip-side stands the argu-ment that major rents from bad regulation or ahighly unfair distribution of rents from naturalresources can lead to economic stagnation aswell as to popular discontent—which ultimatelycan also pose threats to political stability.

Figuring out the nature and implication ofrents and their dynamics over time can be veryimportant for defining country and sector strate-gies. PGPE analysis can focus on diagnosingexisting sources as well as mechanisms for allo-cating rents and the distribution’s relative sta-bility. The mechanisms for deciding on alloca-tions and the actual distribution of similar typesof rents can vary widely across countries. For ex-ample, in some countries, rents from natural re-sources are mainly shared between externalstakeholders and a few key power holders incountry. In other countries, rents are distributedmore widely to balance the political support ofvarious groups and regions.

In terms of evidencing PGPE analysis, it isimportant to note that while it is crucial andusually possible to characterize rents and theirdistribution and dynamics in broad terms, it canbe quite challenging to quantify the size of rents



6. In some cases, stakeholders may deliberately create regulatory rents as part of a development policy (e.g. when pursuing in-fant industry protection). A more pernicious form of regulatory rents are those specifically created to enrich well-connectedindividuals or cliques, without any intention at promoting wider development goals. Regulatory rents created for developmentpurposes can degenerate into the latter form if not effectively managed. See Chibber 2003.

7. Another type of quasi-rents is foreign aid, in the sense that aid is a transfer which did not require a prior effort of local stake-holders (Auty 2007). Using aid as a source of rent-seeking is particularly pernicious when it takes the form of loans which haveto be repaid ex post by all taxpayers.

8. This may involve some important trade-offs as such whether/to what degree to use revenues from natural resources for socialprotection/welfare state spending—which cannot be further addressed here.

or their distribution in detail. For example,property rights over natural resources can bemurky or difficult to trace, or there may be lit-tle public/verifiable information on who bene-fits how from regulatory rents. It can be usefulto reflect up front what level of detail aboutrents and rent distribution needs to be knowngiven the operational question that the PGPEanalysis is seeking to address.

Further reading: Khan and Jomo (2000), Haber,Razo and Maurer (2003), Basedau and Lacher(2006), North (2007).

I.5. PATRONAGE NETWORKS,CLIENTELISM, AND NEO-PATRIMONIALISM

The terms neo-patrimonialism, patronage orpatronage networks, and clientelism have gainedincreasing currency in the analysis of the polit-ical economy of developing countries. In a nut-shell, these terms are used to describe situationsin which formal and informal institutions(strongly) diverge, and informal rules of thegame are subversive of formal ones (see also Box8 above on formal and informal institutions).

Patronage networks and clientelism arewidely found in many countries with varyingimpacts on formal institutions. Neo-patrimoni-alism is the most extreme form of divergence be-tween formal and informal institutions, in whichformal institutions are largely reduced to ‘fa-cades’ (see e.g. Rakner et al. 2004). Many coun-tries with extensive patronage networks andclientelism, especially in Asia, have performedwell with regard to growth and even poverty re-duction; while the problem of such networksstunting development has been most evident inAfrica. Thus, the actual behavior of such net-

works and their impact may depend significantlyon wider sets of incentives. For example, in nat-ural resource rich environments, patronage net-works may have greater incentives to focus onpredation, while in countries with few such re-sources and entrepreneurial opportunities, sim-ilar networks may behave rather differently.

In many settings, political parties as well asethnic and/or regional groups form the organi-zational bases for patronage networks. If pa-tronage networks are dominant at the society-wide level, access to public goods and services,as well as to economic opportunities depends onbeing part of such networks. Receiving a civilservice job, timely access to a good doctor, theprocessing of applications for a business license,or being connected to the electricity grid can allbe determined whether or not a person is part ofa relevant network. At the same time, politiciansand bureaucrats in highly patronage driven so-cieties can be inundated with requests for direct,personal favors.

In other countries, patronage networks aremost important for distributing rents—e.g. al-locating lucrative plots of land in down-town lo-cations or allocating major public procurementcontracts. In such ‘intermediate’ situations, pa-tronage networks are focused on rents, while for-mal institutions play a more effective role, andsignificant parts of service provision and otherday-to-day interactions between citizens andthe state are broadly governed by formal rules.

In neopatrimonial situations, patronage net-works and the informal rules which they createand perpetuate largely undermine and hollowout existing formal institutions. The challenge isthat neo-patrimonialism has a tendency to per-sist because formal institutions are weak andfrequently fail to deliver public goods to citizen;i.e. it can constitute a ‘bad equilibrium’.9

A N N E X 1


9. A key problem of the concept is that its boundaries have remained somewhat poorly defined, i.e. there are no hard and fastcriteria for categorizing a country as ‘neo-patrimonial’.

The concepts of neo-patrimonialism and pa-tron-client networks can be useful to capture thefact that many forms of poor governance are sys-tematic, rather than limited to specific stake-holders or issues. This includes problems of per-vasive corruption. However, governance andpolitical economy analyses should focus on teas-ing out some of the specifics of these structuresand their impacts—rather than just offering avery broad and perhaps superficial diagnosticthat a country is ‘neo-patrimonial’ (see also Erd-mann and Engle 2006).

Furthermore, even full-blown neo-patrimo-nial situations are not static, and change and pol-icy reforms are possible even in such situations.Structural changes of society such as urbaniza-tion and the spread of education are forces thatgradually change social expectations and inter-actions. However, in countries where informalinstitutions and practices are deeply subversiveof formal rules, efforts at reform are likely to re-quire particularly careful attention regardingunintended consequences, potential perversionsof reform, the real incentives affecting the be-havior of stakeholders, and so on.

Further reading: O’Neil (2007); Erdmann andEngle (2006); Chabal and Daloz (1999).

I.6. SELECTED KEY CONCEPTS FROMINSTITUTIONAL AND POLITICALGAME THEORY

Game theory has evolved in many guises acrossthe social sciences since the 1940s. A few keyideas and concepts are briefly surveyed here, asthey may be particularly useful for structuringthe thinking about the political economy situa-tions in client countries.

• Path dependency. Path dependency refers tothe idea that what has happened earlier can

have lasting effects on subsequent situationsand the range of choices. For example,changing or creating agencies now can cre-ate path dependency in terms of the subse-quent range of choices about the publicadministration’s structure. The nature ofcolonial legacies (e.g. Francophone versusAnglophone) created important ‘macro-level’path dependencies in many countries. Im-portantly, however, as mentioned above, for-mal institutions tend to be weaker indeveloping countries than in developed ones,which inter alia makes sudden changes, i.e.ruptures in path dependency, more likelycompared to more stable developed coun-tries.

• Principal-agent relationships and accountabil-ity. The ‘principal-agent problem’ is theproblem of motivating one party to act onbehalf of another. At country level, the chal-lenge is for citizens (as principals) to moti-vate politicians (as agents) to act in thepublic interest, rather than their own self-in-terest or the interest of a narrow group.Within the public administration, elected of-ficials are principals, who need to motivatepublic servants to actually implementadopted policies and to provide services. Thequestion is therefore, what tools principalshave at their disposal to hold the respectiveagents to account. For such accountability towork, ensuring good information on the sideof the principal is assumed to be important;while another important dimension is thepower to act and to enforce accountability(on principal-agent relationships in servicedelivery see World Bank 2004: chapter 3—“The Framework for Service Provision”).

• Collective action challenges/dilemmas (andtragedies of the commons). The notion of col-lective action problems is based on the ob-servation that large groups—such as the



poor, consumers, or owners of small busi-nesses—are often surprisingly powerless, de-spite their large numbers. The concept ofcollective action challenges suggests thatlarge groups often do not organize effectivelyto act in their shared interests. Smallergroups with high stakes have a higher likeli-hood of organizing effectively, potentially tothe detriment of larger groups. A tragedy ofthe commons arises when a shared resource(such as open access grazing grounds orwater sources) are overused, because each in-dividual faces an incentive for overuse, un-less those sharing the commons caneffectively organize to manage it.

• Veto players and gatekeepers. The issue of vetoplayers and veto points as well as of gate-keepers can be relevant for understandingpolicy processes. Legislative processes—andthe ways in which they are formally definedas well as actually used—determine who canblock policy initiatives and at what stages.When supporting policy reforms, it can beimportant to think ahead about potentialveto players and veto points. Gatekeepers inturn are those who channel and select amongmultiple options. For example, ministries offinance are often important gatekeepers withregard to spending requests from line min-istries. At the same time, how a ministry offinance actually performs this gatekeeping

role can be very important—e.g. whether itselects spending proposals based on its meritsfor fostering development, or rather based onpolitical patronage and power considerations.

• Selectorate theory. The selectorate theory fo-cuses on the size of ‘selectorates’ (people in apolity with a voice in selecting leaders) and‘winning coalitions’ (the portion of the se-lectorate sufficient to choose and sustain aleader in office) and the effects of these onthe provision of public goods versus privatebenefits. Democratic systems are assumed tohave the largest selectorates and winningcoalitions, one-party autocracies may havelarge selectorates but have smaller winningcoalitions, and in military juntas, selectoratesand winning coalitions are the smallest. Allpolities are assumed to provide both publicgoods and private benefits, but the larger thewinning coalition is, the more there is an in-centive to provide public goods.

These are just five concepts out of a muchwider range, which have been selected becausethey may be useful for framing particular prob-lems or challenges encountered in client coun-tries.

Further reading: Pierson (2004), Besley (2006),Olson (1965), Ostrom (1990), Tsebelis (2002);de Mesquita et al. (2003).

A N N E X 1


This annex provides further detail on how GPEanalysis can be approached at different levels—country, sector or theme, and project—as set outin Part Three of the main text.

II.1. ‘POLITICS IN ACTION’ AT THECOUNTRY LEVEL

The main value of country-level analysis is to‘cover the ground’—i.e. summarize the countrypanorama. Country analysis is often done as abackground paper (e.g. for the CGAC pilots)that accompanies analysis of more specific issues.Importantly, country-level GPE analysis shouldseek to bring out key political economy dy-namics, rather than providing a static pictureonly. Exploring such dynamics includes a retro-spective of country legacies and how they areshaping the current political and economic sit-uation. While providing a broad overview, coun-try-level analysis can be focused on specific ques-tions or issues, including a first cut at howcountry-level dynamics affect sectors of partic-ular interest. Some options for problem-drivenanalysis at the country level are set out below.

Across the whole range of clients, country-level analysis is valuable especially in the contextof strategic planning, and where there is littleprior GPE analysis. It is particularly crucial insettings where political stability is a concern,since this will often have profound effects acrosspolicy areas and sectors. Country analysis is alsoneeded when major challenges arise, such aspost-election violence and/or clamp-downs,which may require a review of the Bank’s port-folio and assistance modalities.

A vital role of country-level analysis is to de-fine cross-cutting drivers and dynamics. Broadly,country-level analysis covers the following:

i. Foundational dimensions such as the roleof religion, relationships between different(ethnic) groups, but also the role of exter-nal forces and the nature of the politicalregime which form an important backdropagainst which development policies have tobe understood.

ii. Historical factors, capturing in a focusedway how these influence the current politi-cal and economic landscape of the country.1


Annex 2: PGPE Analysis at the Country, Sector/Thematic,

and Project Levels

1. Countries may differ substantially in the degree of ‘path-dependency’, i.e. the degree to which the past has a clear imprint onthe present. Many developing countries have experienced sharp departures, such as transitions to democracy and/or from so-cialism to market economies. However, understanding how the current situation has evolved over time, is an important ele-ment of grasping governance and political economy dynamics.

While legacies are not determining, a senseof how a governance and political economysituation has evolved is important as pastevents shape current relationships betweenstakeholders as well as the current institu-tional set up. Understanding legacies is es-sential to understanding country dynamics,and can highlight the roots of current is-sues, such as variation in terms of the will-ingness of reform across different sectors orthe degree to which past instability or civilwar informs citizens’ attitudes and expec-tations today.

iii. In countries that are affected by conflict orserious inter-ethnic/inter-regional tensions,it is essential to gain an understanding howthe tensions have evolved over time, and totease out key implications for World Bank

programming (such as whether and how tooperate in certain regions).

iv. A map of key institutional structures—including the constitutional system and ac-tual functioning of checks and balances, aswell as key features of the public sector andthe state’s sources of revenue; including amacro-level discussion of the formal as wellas the actual ‘rules of the game’ (e.g. the de-gree to which clientism pervades publiclife).2

v. The main stakeholders relevant for coun-try-level analysis: key political and eco-nomic power-holders; as well as crucialgroups such as political parties, business as-sociations, unions, or religious groups (de-pending on country context and the focusof the analysis).

A N N E X 2


B o x 1 1 : South Sudan GAC for CAS background note

A central strategic objective of the World Bank in Southern Sudan has been to help strengthen corepublic sector governance, including in areas such as PFM, civil service management,decentralization, and anti-corruption. Given the need to build these systems from the previousabsence of an effective state, in the context of a continued politically complex environment, thecountry team engaged in a analysis of the deeper dynamics around the establishment of a publicsector. The process also included an enhanced process of dialogue with government counterpartsat the South Sudan and state level as to potential practical reform options.

The South Sudan Background Note (2008) addresses the key issue that the SPLM/A (SudanPeoples Liberation Movement/Army) has not really yet transformed itself from a fighting forceinto a political movement with a capacity to assume government functions. Continued securityrisks—due to the fact that many groups in South Sudan remain armed—add to this challenge, anddistract from development-focused policy making. Roughly half of South Sudan’s budget isdevoted to security and defense expenditures. On the positive side, there is no imminent risk of acollapse of the Government of South Sudan, as stakeholders fear the consequences of a break-down of SPLM unity, as for the time being people appear to accept the prioritizing of securityover development, and as the international community fills some of the ‘development gaps’.

The challenge for the World Bank and other development partners is to provide the urgentlyneeded support for impoverished and under-developed South Sudan, while managing the risk ofcontributing to uneven development among factions and tribes, and avoiding the emergence ofexcessive dependency. While challenging, the GAC for CAS analysis is considered important by thecountry team as it develops its strategic approach to South Sudan. It also highlighted the need forBank staff to consider the prevalence of security concerns during transitional and post-conflictphases, such as currently existing in South Sudan.

Source: Authors and Motoky Hayakawa.

2. The Dutch SGACA framework can be a useful input to guide country-level analysis. http://www. clingendael.nl/publications/2007/20071000_cru_occ_unsworth.pdf.

vi. An analysis of rents and how rents haveevolved and shifted over time can be highlyinformative about the systemic nature ofstakeholder interests, how they evolved,and how they are linked to wider structuralforces (e.g. volatility in export markets,population trends, etc.).

The analytic lenses outlined in Annex 1 areall pertinent for country-level analysis. Box 12provides a detailed set of questions for country-level analysis which were developed for the GACfor CAS pilots in FY2008, providing usefulguidance as to which issues specifically to ex-plore. Beyond such a mapping of macro-politi-

P G P E A n a l y s i s a t t h e C o u n t r y, S e c t o r / T h e m a t i c , a n d P r o j e c t L e v e l s


B o x 1 2 : Country-level analytical questions developed for the GAC for CAS pilots

Taking the three priorities proposed by the country team as a starting point, the analysis shouldrespond to the following seven questions. (Indicative page limits in parentheses.)

i. How does the country compare with others with respect to the provision of “good policies”generally and in the priority areas identified? Summarize the overall governance profilethrough the ‘Governance at a Glance’ tool. (3 pages)

ii. The credibility and legitimacy of government (vis-à-vis both elites and the broader society) isan integral aspect of governance: lack of credibility and legitimacy reduces the ability of eitherpublic or private actors to plan with a long-term horizon, and makes it more difficult toconstruct a durable coalition that favors the CAS priorities. Both credibility and legitimacy areunderpinned by the allocation of resources and rents. Are agreements regarding rentextraction and distribution accepted by a critical mass of elite interests, stable and credible?That is, do beneficiaries of rent seeking arrangements (whether private investors or keypolitical actors or social groups) have confidence that these arrangements will persist into thefuture? Is there similar confidence in the credibility of property rights? What is the basis of thisconfidence? (15 pages)

iii. One key aspect of governance relates to the incentives of politicians to pursue policies in thebroad public interest. What aspects of the political environment augment or diminish thepolitical incentives to pursue such policies (e.g., those described by the three CAS priorities?)(15 pages)

iv. Another key aspect of governance is the ability of the public administration to carry outpolicies in the broad public interest predictably and sustainably. What limitations doesbureaucratic capacity impose on the ability of politicians to promise and to deliver on theirpromises regarding policies such as the three CAS priorities? Alternatively, what is thepotential of the bureaucracy to implement CAS priorities “technocratically” and sustainably?(5 pages)

v. The pursuit of better governance generally and the three CAS priorities specifically could beextremely risky in countries in which social peace and state institutions are fragile and proneto dissolution from violence. What are the sources of such fragility in the country? Whatarrangements—economic and non-economic mechanisms—seem to be essential to assuringstability—for example, in mediating among deeply divided social groups in the society? Whatimpact would the proposed priorities have on these mechanisms or social equilibria? (10 pages)

vi. The foregoing identified the governance and political economy obstacles and opportunitiesthat affect policies like the three CAS priorities. What is their likely specific applicability to theCAS priorities? How might one address the three development priorities selected for the CASin light of these obstacles and opportunities? (5 pages)

vii. Given the foregoing, are the specific development priorities identified by the country teamthe right ones? Are they the right set of things to pursue given the context? What additionalones are important to consider? (5 pages)

Note: The full set of detailed sub-questions can be found at: http://connect.worldbank.org/units/prem/PD-GPEA/PGPE/default.aspx.

cal features, country-level analysis may includea look at two aspects which often have cross-cut-ting relevance: electoral incentives, and the wayin which policy processes operate (see below).

Especially if done as a self-standing piece, itcan be useful if country-level analysis combinesan overall picture with a focus around particu-lar problem(s) or issues. There are various op-tions of bringing a problem-driven perspectiveto country-level analysis and of giving suchanalysis a clear focus:

i. Summarizing available governance indica-tors and comparing them to regional and/orincome-group averages. Such a mapping ofindicators is useful to stake out up fronthow a country’s overall governance com-pares to peer group(s).3 It can also map thequality of governance against the growthand poverty reduction performance.4

Again, such an up-front assessment helps tofocus the subsequent analysis of governancearrangements and of political economydrivers of these.

ii. Providing a first cut of sectors that are of keyinterest; and exploring what country-leveldynamics imply for those sectors—for ex-ample, strong rural-urban divisions mayhave important implications for agricul-tural policies, or ethnic divisions may havecrucial implications for social protection.This can help to guide subsequent analysis,and to explore how country-level dynamicsappear to affect the more specific politicaleconomy dynamics of sectors. E.g., in some

countries most sectors may be performingpoorly across the board, while in other coun-tries, there is much more variation, some ofwhich may originate in how country-leveldrivers come to bear on different sectors.

iii. Setting out key development challenges fac-ing a country (e.g. improving the manage-ment of HIV/AIDS; managing rapid ur-banization) and providing a broad diagnosiswhether country-level political economyconditions and drivers are conducive to ad-dressing such challenges. A brief summaryof key development challenges and country-wide conditions for addressing them canhelp to focus subsequent PGPE analysis ofspecific issues.

Note that self-standing country-level analysisnecessarily has limits. Where in-depth analysis ofa theme or sector is being sought to inform aCAS or a DPL, it is essential to complementcountry-level work with GPE analysis at sectorand/or project level. Only in selected cases—such as country-level analysis done in reaction tocrisis and with the purpose of informing re-programming—will country-level GPE analysisbe directly operationally relevant. That said, itoften has major effects on how teams thinkabout country context, and hence it can havesignificant indirect relevance. Also, it is highlyvaluable in establishing a context for more issuespecific GPE analysis.

For an increasing number of countries, someform of country-level analysis has already beenconducted.5 Where ample material is available,

A N N E X 2


3. Note that there are (still) issues with the precision and comparability of governance indicators; but they can provide a useful‘frame’.

4. E.g. Benin would be a country with relatively poor governance on several dimensions and relatively low growth and povertyreduction performance; while Cambodia combines poor governance performance with high growth and significant povertyreduction in recent years. Mapping this can help to frame the subsequent analysis.

5. In particular by DfID and the Netherlands, as discussed in section 1. However, it is worth noting that there is not yet an es-tablished practice of sharing governance assessments across donors. E.g. the Netherlands has not yet shared any of its SGACAassessments.

it may suffice to do country-level analysis as abrief summary combined with an update onmost recent events (e.g. new election, majornew policy initiatives; corruption scandals,etc.) and their implications. In many cases itis still highly valuable to organize, analyze,and summarize existing material in a way thatmakes it accessible and pertinent for countryteams and their understanding of the politicaleconomy context of the country. Where no rel-evant country-level analysis has been done,country-level analysis will require more ex-tensive work.6

Country-level analysis is the only one of thethree levels which can potentially be done purelyas a desk-based exercise, especially if ample ma-terial exists. However, for a more intense inter-action with country teams to define the issues tobe addressed, and also to get a high-quality up-date of current events, experience shows that amission adds significant value (see Part Five,sections 5.1 and 5.2).

II.1.1. Special module 1: Analyzingelectoral incentives and theirpotential importance for policy andpublic investment choices

While the World Bank has no involvement in re-forming electoral rules, electoral incentives oftenexert considerable influence on how policies areshaped as well as on public investment choices,and hence this has become a specific themewhich is beginning to be explored as a module

or section of country-level analysis. Under-standing these incentives from a diagnostic per-spective can be of considerable benefit to Bankteams working on policy and public sector re-forms.7

There are four key dimensions: (i) electoralrules and the incentives they create, (ii) the wayin which electoral promises are made, (iii) the de-gree of information available to voters, and (iv)the crucial factors shaping voter preferences.8

i. The first dimension directly affects howcandidates and MPs behave. For example,majoritarian electoral systems—in whichcandidates are elected from one particularlocality—often create incentives for local-ized public spending. Electoral rules canalso affect the degree of concentration ver-sus diffusion of power. While some elec-toral rules are clearly flawed, no set of rulesis automatically superior to another (see Coxand McCubbins 2001; Fukuyama 2007).9

ii. In some settings, a discussion of policies isalmost absent from election campaigns,which become almost entirely focused onpersonalities and on clientelistic promises,including promises of ‘bringing’ roads orschool buildings, or the like.10 Alterna-tively, some parties and politicians makehighly populist policy promises—especiallyif they need to make a nation-wide ratherthan local appeal in order to win elections.Generally, few developing countries havepolitical parties that have a well-defined



6. If affordable, country visits are almost always valuable as they establish a closer interaction between the PGPE analyst and thecountry team; and because they are invaluable for capturing most recent trends.

7. Such analysis can only usefully be applied in countries with at least reasonably free and competitive elections.8. Analysis can focus on presidential or parliamentary elections or both, depending on country circumstances.9. Frequently, problems rest with secondary rules, such as the fact that in a majoritarian, localized system, politicians are not re-

quired to actually reside in the district they are being elected from, undermining the localized accountability which the sys-tem as such implies.

10. Also, in countries in which identity politics are highly important (i.e. where people tend to vote along ethnic, religious or lin-guistic lines), voters may be more distracted from voting about policy issues, than in places that are more homogenous interms of people’s identities.

programmatic base (see Randall 2007).This implies that their policy agendas areoften highly opportunistic and changeable.

iii. The amount and quality of informationthat voters have available are driven by gen-eral country characteristics, such as the levelof education and urbanization, the politicalregime, as well as by more specific condi-tions such as budget transparency, the exis-tence of M&E systems, and media freedomand (investigative) capacity.

iv. Regarding the fourth dimension, votersmay be motivated by different prefer-ences—not only their own economic self-interests, but also regional or group

loyalties. In many settings, voters preferwell-known individuals, which may in-clude sportsmen or film stars.

Overall, understanding electoral incentivescan be highly valuable in understanding whatkinds of policies are politically attractive (or un-attractive), the politics of allocating public in-vestments, as well as the overall nature of ac-countability between society and political elitesand the effectiveness of demands for good gov-ernance and development-oriented policies. Indecentralized settings, and analysis of electoralincentives can also contribute to understandingmore local accountability relationships.

A N N E X 2


B o x 1 3 : Electoral incentives in Benin and the lack of growth-focused policies

Over the past two decades, many low income countries have begun to establish democraticsystems, and have been holding more or less free elections. The puzzle is, however, that eventhough democratization in principle gives citizens the ability to choose their government, in manycases this has not translated into governments delivering policies aimed at growth, povertyreduction, and overall development—which would benefit citizens at large (see also Keefer andKhemani 2005).

The Benin CEM analyzed electoral incentives in order to understand why this link has beenweak. The analysis highlights that in a poor, young democracy such as Benin, citizens often havelimited knowledge and information on what government policy can do, how it affects economicgrowth, and how economic growth affects their own welfare. In turn, politicians who perceivelow political payoffs to pursuing growth-accelerating reforms are less likely to undertake them.

The competition for votes is focused around clientelism rather than policy promises. Politicalcompetitors are unable to make credible pre-electoral promises regarding growth-related policies,as they lack the ability to implement such policies even if elected.

The electoral system in Benin is proportional (i.e. based on party lists), divided into a numberof election districts (rather than one nation-wide list). The system encourages a proliferation ofparties; while the fractionalization of parliament means that individual or small groups of MPshave little chance of promoting any larger-scale policies. The focus is therefore on personalpopularity and trying to provide some direct benefit to election districts.

In public opinion surveys, 95 percent of respondents said that politicians rarely or never keeptheir election promises (compared to an average of 82 percent for all African countries). ElectedMPs are expected to deliver individualized benefits, and usually receive a slew of requests tocontribute to marriages, naming ceremonies, burials, education fees and health bills, while votersdo not expect them necessarily to provide public goods.

Furthermore, even though the press is relatively free, it is difficult for Beninois to obtainrelevant information on government policies regarding growth and government service delivery;as few are able to afford newspapers, and as radio stations often provide little information onpolicy issues. Educational levels in Benin continue to be low. Consequently, better education andinformation could be a key to building greater demand around development-promoting policies.

Source: World Bank 2008a, Chapter 2.

II.1.2. Special module 2: Analyzing policy processes

Analyzing policy processes can be fruitful forcountry-level as well as for thematic or sector-focused GPE. Some key generic qualities ofpolicies and policy processes are: public-regard-edness (the degree to which policy making is inthe general public interest, rather than the in-terest of specific individuals or groups), policystability and adaptability, policy coherence andcoordination across policy areas, and the qualityof implementation and enforcement (see IADB2006: 17). In general, policy stability and pre-dictability is important, and policies should notbe changed randomly, e.g. driven by changes inthe ‘political weather’. At the same time, coun-try systems need to be able to produce or adaptpolicies when needed—because policy gaps be-come apparent, existing policies fail, or becauseexternal circumstances change. Among the driv-ers that are decisive for the quality of policies areelectoral incentives (discussed above), the formalconstitutional set-up as well as more specific

rules determining the role of key players in theexecutive and legislature and the structure ofveto-points, as well as the nature of politicalparties and the political culture more broadly.Last but not least, aid dependency and the roleof donors and international agencies in shapingpolicies can also be an important factor.

Analyzing policy-making processes can clar-ify the formal and de facto mechanisms fordrafting and adopting policies. This can be donein a more generic form (e.g. teasing out generalmechanisms and patterns), or through processtracing of one or several specific policies that areof particular interest (see also World Bank2007c). In a number of instances, World Bankteams have been surprised by policy decisions; orconversely by failures of policies to be adoptedeven though they had been agreed and appar-ently supported by governments. Lack of im-plementation subsequently to adopting policiesis also a frequent problem in many countries.Analyzing policy processes and what drives themcan help to reduce surprises. Such analysis mayenable teams to tailor their policy dialogue and



B o x 1 4 : Analyzing policy-making processes in Latin America

The IADB has undertaken a major research effort to understand public policy making and its driversacross the region. Key findings are published in IADB (2006) and in Stein and Tommasi (2008). Oneexample is an analysis of the policy making process in Paraguay. In Paraguay, most policies areinitiated by the executive, and some by the legislature. There appears to be also a division in policycontent: the executive tends to dominate policy proposals with nation-wide scope, while thelegislature initiates more bills targeted at particular localities, but also at the private sector. Theshare of bills passed compared to bills introduced declined from the early 1990s to the early 2000s.The analysis also explores whether bills are ‘particularistic’ i.e. benefiting certain groups (‘pork-barrel’) or non-particularistic. Particularistic bills have generally been adopted more easily, andhave also faced fewer problems during implementation, while non-particularistic policies havestalled more frequently during adoption or implementation. Parliament has become relativelyfragmented over time, and as a consequence has often acted as a veto-player, blocking important(non-particularistic) policies such as reforms in the telecommunications sector, social security, thehealth system, the civil service, and financial sector regulation. The fact that particularistic policiesare less controversial and more easily passed means that the public-regardedness of policy makinghas suffered.

Source: José Molinas et al. in Stein and Tommasi 2008, pp. 329–369.

advice more effectively, in terms of timing, in-terlocutors, as well as the type of arguments andapproaches (e.g. moving beyond a predomi-nantly technical approach which can only reachcertain audiences). Importantly, in each countrythere are general characteristics of policyprocesses (such as the predominance of the pres-ident in initiating policies), as well as featureswhich are specific to various policy areas due tothe fact that different configurations of stake-holders and interests are concerned, dependingon the policy area (IADB 2006: 21).

There are different elements of analyzing pol-icy processes: one is to map the existing formalmechanisms and how they are used de facto—e.g. which individuals or group can submit pol-icy proposals, what are the prescribed stages fordiscussing policies (in cabinet, in parliament),and to what degree and how does the public orinterested groups such as the business commu-nity need to be consulted. Regarding imple-mentation, it may be possible to track whetherkey policies—such as taxation, trade policies,etc.—tend to be universally implemented andenforced, including whether implementing reg-ulations tend to be in place, whether policies areactually budgeted for (including funding forimplementing institutions) and so on. For spe-cific policies it can be useful to establish howthey were proposed, discussed, adopted, andsubsequently implemented. Minutes of parlia-mentary discussions, as well as interviews withstakeholders who participated in the process,can bring to light how support for a policyemerged, or even why certain provisions were in-cluded or omitted.

II.2 SECTOR-FOCUSED ANDTHEMATIC PROBLEM-DRIVEN GPE ANALYSIS

This section discusses sector-focused and the-matic GPE analysis. This level is highly perti-nent for Bank policy and lending operations.First, the section considers sector-focused GPEanalysis, and addresses specifically GPE analysisfor public sector (or ‘core governance’) reforms,and analysis for service delivery sectors morebroadly. Section II.2.2 then turns to thematicGPE analysis. For example, at this level, GPEanalysis may focus on thematic questions ofpoverty reduction, growth issues, natural re-source management, or pollution and climatechange adaptation.

II.2.1 Sector-focused PGPE analysis

Sector operations are at the heart of the WorldBank’s work. However, sector policy dialogueand sector operations have repeatedly run intodifficulties because of a lack of thorough un-derstanding of the political economy environ-ment. Problem-driven GPE analysis upstream ofsector operations and/or during their imple-mentation can make important contributions tobetter design and implementation management.

This section aims to provide some generalprinciples of how to approach sector-focusedGPE analysis. Some tools for assessing gover-nance and corruption at sector level are alreadyavailable (e.g. PETS for expenditure tracking inhealth and education, and PEFA for PFM).11

However, most of these tools do not drill down

A N N E X 2


11. See for example: http://go.worldbank.org/HSQUS 4IS20 (PETS—WB intranet only), www.pefa.org (PEFA), http://go.world-bank.org/VZ476DL7P0 (Improving governance in the electricity sector). Existing governance surveys and other sector-focusedgovernance diagnostics are a valuable input to GPE analysis, as they should provide key information for diagnostic layer 2 (‘ex-isting governance arrangements’). To date, detailed governance sector diagnostics have only been carried out for a small num-ber of pilots, but the rolling out of GAC activities should increase such applications.

to the political economy drivers. Where such as-sessments are available, they can be very helpfulin providing a good understanding of the firstand especially the second diagnostic layer (gov-ernance arrangements) as set out in part 1, whileproblem-driven GPE analysis as described herecan then be used to delve more directly into ad-dressing underlying political economy drivers.

This section first discusses sector reforms fo-cused on ‘core governance’—i.e. public sector re-forms—that are at the heart of overall gover-nance in a country. It then briefly sets out somekey issues for GPE analysis for other types of sec-tors. It is expected that more specific guidancewill emerge over the course of FY09–10 tai-lored to the diverse range of sectors.

II.2.1.1. Problem-driven GPE diagnosticsfor core governance reforms

As is the case for other areas, public sector re-forms have predominantly been pursued through(exclusively) technical approaches. For the WorldBank’s mission, the most important core gover-nance areas are public financial management,the civil/public service, anti-corruption, and thejudicial system.12 Decentralization can be partof core governance reforms depending on coun-try circumstances. While many operational staffimplementing and supervising public sectorprojects are broadly aware of political economyissues, this mostly remains ‘tacit knowledge’.While still valuable, it means that there is littleexplicit discussion of political economy factorsand usually little effort to define how these af-fect operations and how they might be ad-dressed.

IEG recently evaluated the effectiveness ofWorld Bank support for public sector reforms

(IEG 2008). The evaluation concludes thatwhile there are many areas of success, a signifi-cant number of interventions have not broughtexpected results. Furthermore, the report arguesthat largely leaving political economy factorsunaddressed is one of the key reasons why de-velopment impact has not been more consistent(IEG 2008: 73–74).

Crucially, the political economy of publicsector reforms tends to be especially closely con-nected to macro- or country-level political econ-omy issues. At the same time, public sector re-forms also have ‘internal’ political economydynamics (which are sometimes referred to asthe ‘micro-political economy’ of public sector re-forms) (see Figure 5).

In order to get a good grasp of possible entrypoints, strategies and opportunities for reform, itis important understand with some precisionhow macro-level political economy factors in-teract with those at the micro level. For example,civil service reforms are closely linked to how thepolitical layer of the state operates, such as thebalance between patronage and clientelism ver-sus a performance orientation, or the ability ofcitizens to voice demand for better services. Atthe ‘micro-political’ level, there are often specificpolitical economy dynamics, for examplearound the roles of civil service commissions, orconcerning the functional re-ordering of de-partments. An important ‘connecting point’ isthe organizations, institutions, and incentiveslinking the country-level political economy withthe different areas of public sector reforms. ForPFM, for example, ministries of finance are im-portant organizational stakeholders. In contrast,there is no equivalent stakeholder in most coun-tries for civil service reforms. The relative polit-ical clout of ministries of finance can allow min-



12. Other development partners include democratization and human rights as part of their core governance agenda, but these arenot areas that the Bank actively engages or has a comparative advantage in.

istry officials to pursue certain technical reforms,even if these reduce rent-seeking opportunitiesfor other stakeholders. At the same time, evenministries of finance may often not have theclout to overhaul procurement, or to entirely re-design budget processes.

Some teams have started to pilot GPE-typeanalysis for public sector reform projects, andfurther work is being planned by PRMPS as wellas PREM regional units to fill this gap.13 Box 15reflects some of the early examples.

II.2.1.2. Problem-driven GPE analysis forservice delivery sectors

Sector reforms—from social sectors, to agricul-ture, to electricity, roads, and water—are crucial

to achieving poverty reduction and growth. Sim-ilar as for public sector reforms, evaluations ofWorld Bank operations have pointed to theneed to take the political economy context of re-forms into account more explicitly in order toimprove development effectiveness (see IEG2006). Through the PSIA approach, pilot workhas been done, especially on the political econ-omy of reform in agriculture and in water sup-ply and sanitation (World Bank 2008b). Prob-lem-driven GPE analysis of sectors has also beenpiloted in the context of CGAC and sectorGAC processes, e.g. in Mali, Zambia, andLebanon, with a focus on infrastructure sectors,and especially energy.

sector-focused GPE analysis may be done aspart of a CAS process, motivated by an experi-

A N N E X 2


F i g u re 5 : Macro and micro political economy interactions in core governance reforms

Judicial reformCS reformsPFM reforms Anti-corruption

Macro- / country-level governance and political economy

Reform specific political economy dynamics Interactions between different

public sector reforms

13. E.g. PRMPS has commenced a series of pilots on the capabilities of Ministries of Finance, including an assessment of howthis is impacted by political economy factors.



B o x 1 5 : Governance and political economy analysis informing public sector reformoperations—Moldova and Afghanistan

Moldova initiated an ambitious Public Administration reform in 2005. Key motivations were theneed to improve government policy making and administrative capacity more broadly and tofacilitate closer integration with the EU.

The four explicit goals were: (i) a significant reduction of ‘redundant’ personnel at the centrallevel, (ii) subsequent raise in salaries to competitive levels, (iii) elimination of redundant institutionsthat perform overlapping and outdated functions, and (iv) improvement of capacity to implementgovernment strategic objectives. The program has been supported through a donor trust fund.Milestones of the reform were mapped out in an overall action plan, complemented by annualaction plans.

However, actual reform progress was limited. A governance and political economy analysishelped to explain the key reasons for this. One, the mix of trying to achieve qualitative reformsand of pursuing a significant downsizing of the civil service prevented the emergence of buy-inamong ministries and agencies. The president ultimately also blocked important components ofthe reform, because these would have reduced his direct influence over personnel decisions.Country-level analysis also pointed to the fact that those social groups which would most likelydemand greater state effectiveness and accountability are weakened due to very high rates of labormigration of the working-age population. Demand from NGOs on this issue has been rather weak.

Some partial reform has been possible nonetheless, including the adoption of a new CivilService Law. The EU continues to be a driver of reform, as closer approximation is a political goal,and since at the practical level the EU is promising significantly increased budget support if thepublic financial management aspects of the public administration can be improved.

Afghanistan. A brief note on the Political Context of Public Administration Reform in Afghanistanwas produced in 2006 by a prominent political scientist. The note reaches back into the historicalevolution of a public service in Afghanistan, including the history of earlier external interventions.Crucially, the note shows that the international community (led by the UN) attempted to create asmall, meritocratic core civil service at the outset of state (re-)building in 2001. However, theseideas were not accepted by the various factions—all of which were keen to ensure that they couldappoint loyal members of their factions; as well as to maintain full discretion in appointing civilservants (rather than giving a prominent role in such decisions to a civil service commission).Limited employment opportunities outside the civil service, and the fact that former publicservants—a number of cohorts of which had been created by the changing regimes—in principleall received the right to return to their positions compounded the problems. However, comparedto some other post-conflict countries, the size of the civil service has still remained limited, due toa crucial gate-keeping role on the wage bill exerted by the MOF.

In 2003, an Independent Administrative Reform and Civil Service Commission (IARCSC) wascreated, and a civil service commission was set up under its guidance. Initially, the IARCSC focusedon a program to establish centers of excellence within ministries, with the idea that most valuableofficials would be transferred to Priority Reform and Restructuring (PRR) units and receiveadditional salaries; and that over time ministries would be slimmed down. Success of this approachhas been mixed. Some ministries have begun to pursue greater results orientation throughcontracting with NGOs (e.g. in the health sector). The PRR was subsequently phased out, andattention shifted to the reform of ministries overall.

The World Bank has been engaged in Public Administration Reform both on the PFM and theCivil Service side in Afghanistan since 2002. The projects have continuously been informed by asubstantial body of AAA. The governance and political economy analysis in particular has madeimportant contributions to striking the balance between seeking to re-build a modern publicadministration and being realistic about the country context. This has included defining the overallstrategic approach, and focusing on specific entry points.

Source: Authors and Jana Kunicova (Moldova) and Ranjana Mukherjee (Afghanistan).

ence of only partially successful or unsuccessfulsector reforms in the past, and/or as part of theprocess of preparing sector operations. As men-tioned above, sector-focused GPE analysis canbe done as a package together with country-levelanalysis (e.g. as was done in FY08/09 for Zam-bia and Lebanon). In the same vein as for theother levels discussed here, problem-driven GPEanalysis for sectors should seek to map out chal-lenges or vulnerabilities, governance and insti-tutional arrangements, and based on these drilldown into the political economy drivers.

A sectoral value chain approach (SVCA) canbe helpful for disaggregating a sector and defin-ing what aspects are most important to analyze(see Figure 6).14 For example, the power sectorcomprises the overall policy and enabling envi-ronment, planning, budgeting and financing ofparticular projects, followed by their tenderingand procurement and subsequent construction,system-wide service delivery, operations andmaintenance, and the ability to achieve expan-sion and quality improvements. The educationsector is composed of the different levels of ed-ucation (primary, secondary, tertiary), the fi-nancing of the sector, staffing and teacher train-ing, actual delivery of teaching on the ground(often involving local governments), text-bookprovision, the regulation and supervision of ex-ams, and so on. The capacity and commitmentof governments to develop and enforce effective

regulation on the one hand and the ability of cit-izens to hold service providers to account eithervia long routes or short routes of accountabilityon the other hand cut across all these links of asector value chain.15

In many cases, the problem which GPEanalysis is being asked to address is already setout in ongoing sector analysis. However, it mayalso be the case that the problem (such as poorlyprogressing sector reforms) is not clearly set outin any written document. Also, GPE analysis it-self may bring some ‘neglected’ vulnerabilities tothe fore, such as weak accountability mecha-nisms or specific rent-seeking activities, whichmay have knock-on effects, for example, onpoor grid maintenance or creating incentiveproblems in the management of schools. A firststep is therefore to set out and describe the chal-lenge that is being addressed.

For the governance and political economylayers of the analysis, the task is to tease out thegovernance arrangements and underlying po-litical economy drivers in the sector and, respec-tively, the selected aspects of it. Firstly, this in-volves mapping out the institutional andgovernance arrangements relevant for the se-lected aspects and the formal rules and de factobehavior that shape sector policies and sectormanagement.16 Importantly, problem-drivenGPE analysis at this level should consider weak-nesses of ‘commission’ as well as of ‘omission’. A

A N N E X 2


F i g u re 6 : Using value chains as a tool to disaggregate sector components

Policy and enabling environment

Planning, Budgeting and finance

Tendering and procurement

Construction/installation/supply

Service delivery, operations and maintenance

Expansion, quality improvements

Cross-cutting: accountability & effective government regulation (setting standards, tariffs, etc.)

14. The idea of such sectoral chains draws on Campos and Pradhan (eds.) (2007). 15. The long route refers to the ability of citizens to demand accountability via voicing demands to politicians and policy makers. The

short route refers to citizens’ ability to hold frontline service providers directly to account. See World Bank (2004): chapter 3. 16. Some governance diagnostic work may already exist that can be utilized.

sector may be affected by corruption (a ‘com-mission’ weakness, which many governance di-agnostics focus on), but it may also suffer froma failure of the public sector to take action (an‘omission’ weaknesses). For example, health caremay be poor not only because existing govern-ment programs are affected by corruption, butalso because the government is unable to mustera more concerted effort, and/or is failing to pro-vide better infrastructure which would facili-tate better access to health care.17

Secondly, identifying stakeholders and theirinterests is crucial (see Annex 1). However, cat-egorizations of ‘winners’ and ‘losers’, or ‘reformchampions’, ‘undecided’, and ‘opponents’ needto be done with care and conscious of the factthat positions can shift, and that stakeholders of-ten have interests that cut across several sectorsor issue areas. For example, business associa-tions may have stakes in tax policy as well as fuelsubsidies; while farmers may worry about wa-ter/irrigation policies, as well as health-care and



B o x 1 6 : Sample questions for sector-focused GPE analysis

• What is the ownership structure in the sector? (public & private)• How are responsibilities distributed between the national and sub-national levels and is this

distribution clear? Does it generate significant distortions?• How is the sector regulated (what are the rules and institutional structures)? Does existing

regulation—including the informal/de facto rules—provide integrity? Does existing regulationallow the sector to maintain or expand services in line with demand (and commitments topoverty alleviation)? What interests drive/maintain the current regulatory system (including itsweaknesses or gaps)?

• How are the sector and its components being funded? (e.g. user fees, taxes/general budget,earmarked taxes (such as gasoline excise taxes), informal revenue generation, petty corruptionfrom consumers, etc.)

• What is the pricing structure for consumers? Which groups benefit (e.g. from subsidies)? Arebenefiting groups politically salient/powerful? Which consumer groups have a voice?

• Is there significant petty corruption and/or grand corruption in the sector—and if so, why doescorruption persist and what are the main impacts?

• What opportunities for rent-seeking and patronage are related to the sector? Who appears tobenefit from these rents and how is the patronage being used?

• What are the legacies of the sector? What reforms have been attempted and/or undertaken inthe past? What were the results—and how does this experience appear to shape currentexpectations of stakeholders?

• What are the relevant policy processes linked to past or proposed reforms?• Are there particular social or ethnic factors that are relevant for sector dynamics? • What is public opinion on sector performance and/or proposed sector reforms (including issues

of trust/expectations that a reform would bring improvements)?• What stakeholders are (officially and unofficially) involved in discussions over sector reforms

and what are their interests? What veto points exist in the decision-making and theimplementation process?

• What stake does the government/top executive/key political factions have in the reform if any?• How would proposed reforms affect the existing set of interests and incentives?• What risks exist in terms of reform failure and/or of negative unintended consequences of

proposed reforms?• What would a politically and institutionally feasible reform look like?

Source: Authors.

17. This does not imply that governments would be expected to be the sole or main provider of all these services; however, evenif private or public-private provision is an option, this often requires some effective government action in order to materialize.

roads. Assessing political economy drivers in-cludes the question whether there are significantrents or rent-seeking opportunities and howthese would be affected by proposed reforms.

Furthermore, historical legacies and experi-ence of previous reforms and potentially theirfailure can influence the perceptions and expec-tations of stakeholders, and are therefore poten-tially worth exploring. Tracing particular sector

policy decisions (or proposed decisions that failed)can be helpful for gaining an understanding ofhow stakeholders and institutions interact, both‘horizontally’ within the sector and verticallywith wider country-level institutions such asparliament and various levels of the executive.

As set out in Part One, the aim of PGPEanalysis is to define feasible interventions—andthis is particularly relevant for sector-level analy-

A N N E X 2


B o x 1 7 : Examples of PGPE for service delivery sectors—transport in Bangladesh,electricity in Lebanon

BangladeshThe transport sector is a major sector of the Bangladesh country program. However, there aremajor concerns that the sector is affected by corruption and rent-seeking interests, especiallyrelated to road construction. An Operational Risk Assessment of Roads and Highways Investmentswas carried out in FY2007. Beyond diagnosing the significant extent of corruption in the sector,the assessment drilled down into key political economy drivers—which included a political monopolyof road expenditure decisions, and an established market for public office in areas that are lucrativefor rent-seeking. The sector analysis could also draw on an IGR, which had been undertaken a fewyears earlier, that set out broad country-level governance and political economy issues.

The operational risk assessment recommended that simply putting more controls into placewould be insufficient, as they would either be broken/circumvented, or just be used to shift rentsfrom one group to another. The recommendation was therefore to combine support to relevantgovernment departments with a strengthening of public participation in road-sector decisionmaking, and a public information campaign to raise awareness about proposed reforms and issuesat stake, and strengthening the capacity and visibility of key oversight bodies.

LebanonThe electricity sector in Lebanon has been plagued by problems of high inefficiency, rising numbersof brown-outs and black-outs, and being a growing fiscal drain due to low tariffs and risingsubsidies for the state-owned electricity company. Despite the rising fiscal cost and increasingdiscontent of consumers with poor services, successive governments have not taken action toaddress the problem.

As the sector-focused analysis reveals, in addition to the corruption and rent-seeking problemswhich are typical in many countries, an added dimension in Lebanon is the fact that differentgroups and factions in the country have each sought to secure control over parts of the powersupply, during acute conflict as well as more peaceful periods (local elite capture).

To date, most stakeholders who have some effective voice have jockeyed for privileged accessto the subsidized energy, while ensuring themselves with back-up generators against outages.These include capital city electricity consumers, as well as the main industrial groups. Thus, therehas been little demand for sector reforms.

The assessment suggests that the Bank’s preferred ‘first best’ solution, unbundling andprivatization of the state-owned electricity company (which has already been stipulated in a lawwhich has remained unimplemented for the past six years) may not be feasible. Instead, theproposition is to work on building a broader demand coalition for reforms, focusing onincremental steps such as improving metering to reduce non-technical losses, and possibly to tacklereforms in selected regions first, rather than seeking overall changes at the national level.

Sources: Bangladesh—Operational Risk Assessment of Roads and Highways Investments (2007) and LebanonDraft CGAC Report (2009).

sis. There has often been a tendency to espousesolutions in policy and lending operations andin sector policy dialogue that do not take intoaccount what is feasible given political econ-omy and institutional conditions. As set out inPart Two, sector-level GPE analysis should feedinto a process of deciding on appropriate oper-ational interventions. This includes a consider-ation whether to primarily select what inter-ventions would be feasible given the existingreform space, or whether it is appropriate andfeasible to pursue a more proactive strategy ofseeking to expand the existing reform space.

II.2.2. Thematic problem-driven GPE analysis

In a number of situations, country teams may belooking for GPE analysis around certain themesor issues. The focus of thematic analysis isbroader than in the case of a single sector andthe analysis would usually cut across several sec-tors. A team may be concerned about a series ofpolicy reforms around an issue (such as the po-litical economy of poverty reduction), or it maybe looking more for a background piece.

Themes that have been explored through pi-lot GPE analysis include:

• Managing natural resources/the risk of a re-source curse18

• Complementing growth analysis with prob-lem-driven GPE analysis, notably aroundbinding constraints to growth and defining‘feasible’ reforms to reduce constraints

• The challenges of addressing persistentpoverty effectively.

The basic principles of PGPE analysis can beapplied to each of these themes or issues. Fol-lowing the three diagnostic layers, the work canstart by defining the vulnerabilities or prob-lems, looking at the governance arrangementsand policies related to the issue (which will of-ten span several sectors or levels of government),and then seeking to capture the underlying po-litical economy drivers. Thematic analysis isperhaps the most challenging type of problem-driven GPE analysis, since on the one hand it re-quires more specific evidence than overall coun-try-level diagnostics and pertinent informationmay be less readily available, while on the otherhand it does not offer the benefit of narrowingto a specific sector or policy question, includinghaving a rather clearly defined set of stakehold-ers and institutions concerned. Nonetheless,some good and valuable examples of such workare emerging, as well as specific resources for spe-cific areas of engagement (see below). As set outin the Introduction, this good practice frame-work does not seek to provide detailed guidanceon problem-driven GPE for specific themes.Frameworks for specific themes and issues mayemerge over time, as is already evident for thetheme of natural resource management and forpolitical economy for growth diagnostics.

II.2.2.1. Governance and political economyanalysis for growth analysis

The importance of economic growth for sus-tained poverty reduction has gained increasedprominence among development partners overthe past few years. The growth policy dialogueof the World Bank is often focused on analysisconducted as part of country economic memo-



18. NRM is defined here as a theme rather than a sector, because it comprises not only mining as such, but also specific revenueand expenditure management challenges, environmental protection, and so on. The boundaries between a sector and a the-matic analysis are somewhat open.

randa (CEM). Applied work in the past fewyears has focused on sharpening the prioritiza-tion of key constraints to growth. For example,the Hausman-Rodrik-Velasco (HRV) approachto growth diagnostics sets out a decision tree thathighlights particular types of constraints. Al-though there are other methods of growth di-agnostics, the HRV approach has been increas-ingly applied as part of the World Bank’s CEMs.This ideally yields a set of discreet types ofgrowth constraint problems, ranging from gov-ernment failures (including corruption, lack orrule of law, poor taxation), to infrastructure,human capital, or financial sector constraints.These constraints can impinge largely at thecountry level, or have regional dimensions (in-cluding when countries confront poor/land-locked geography).

Given a range of possible binding constraintsto growth, GPE analysis can useful complementgrowth analysis to help explain the underlyingpolitical economy drivers of these constraints.For an approach to applying GPE to growthanalysis, notably for the case of infrastructure inZambia see Fritz (2008). A number of recentCEMs, including in Benin and Ghana, have al-

ready begun to pay more specific attention ofpolitical economy issues (see Badkoubei andKaiser 2009). Such political economy analysiscan be catered to particular areas of public serv-ice delivery (Keefer and Khemani 2005) and offinancial sector development (Keefer 2007).

Resources: http://connect.worldbank.org/units/prem/PD-GPEA/growth/default.aspx (intranet)

II.2.2.2. Governance and Political EconomyAnalysis for Resource-led Development

The development of settings rich in natural re-sources (oil, gas, mining, forestry and fisheries)continues to present a significant challenge. Dis-appointing historical performance has led anumber of observers to characterize the resourcewealth as a curse rather than blessing. The EI forDevelopment approach (or EITI++) has em-phasized that resource rich countries need tomanage all parts of the value chain, from estab-lishing contracts for mining, to how resourcesare managed, to how they are spent (e.g., in-cluding public investment management) (seeFigure 7). Systematic GPE analysis can provide

A N N E X 2


F i g u re 7 : Sector value chain for natural resource management

AWARD OFCONTRACTSAND LICENSES

NON-RENEWABLERESOURCES

SUSTAINABLEDEVELOPMENT

REGULATIONANDMONITORINGOFOPERATIONS

EITI

COLLECTIONOF TAXESANDROYALTIES

REVENUEDISTRIBUTIONANDMANAGEMENT

IMPLEMEN-TATION OF SUSTAINABLEDEVELOPMENTPOLICIES ANDPROJECTS

implications of how particular segments of theEI value chain could be improved.

The political economy literature has also rec-ognized that natural resource rents can funda-mental affect the nature of the state-society re-lationship. In natural resource settings, this mayalso call for particular political imperatives forstrengthening non-resource based taxation as away of enhancing accountability. Consequently,from a structural perspective (see Annex 1), thepresence of significant natural resources canmatter significantly for the overall macro and‘micro’ political economy of a country.

Resources: http://connect.worldbank.org/units/prem/PD-GPEA/EITIplusplus/EITIhomeWiki/Home.aspx (intranet)

Frequently, thematic GPE analysis will feedinto several processes, including operations inways very similar to the case of sector-focusedanalysis. For example, thematic GPE analysis fo-cused on the overall challenges of natural re-source management may come to inform a min-ing sector operation, as well as potentiallyoperations on PFM. In the case of thematicanalysis focused on GPE dimensions of con-straints to growth, results may feed into theoverall policy dialogue on growth, as well asinto operations related to specific constraintsto growth such as investment lending to improveroads and railroads. Other potential themes aremanaging environmental resources and risks/ad-justing to climate change, managing food secu-rity risks, and so on.

II.3. PROBLEM-DRIVEN GPE FORSPECIFIC PROJECTS OR SINGLEPOLICY DECISIONS

Teams may seek to understand governance andpolitical economy issues directly related to a

specific project or a single policy decision. Thiscan be done as an extension of sector-level analy-sis and/or it can take the form of quick turn-around notes that complement other types ofanalysis done by task teams during projectpreparation or supervision. For investment proj-ects and TA, key questions for GPE analysismay be whom to work with, to understand theincentives of stakeholders better, as well as to as-sess whether the circle of stakeholder a projectengages with should be broadened, for exampleto include demand-side actors, and how thiscould best be done. GPE analysis also con-tributes to a more thorough understanding ofthe institutional landscape that a project inter-acts with—be it the structure of an urban gov-ernment, or a specific government agency, orSOE and its regulators and so on. For specificpolicy decisions that Bank teams engage with inthe context of policy lending, targeted GPE analy-sis offers an opportunity to assess feasibility, as wellas to explore potential unintended consequences.

In some sense, GPE analysis at this level isideal, because it allows a clear focus and the ex-ploration of a very specific set of institutions, in-centives, stakeholders and interests. For example,some stakeholder modeling work has been doneexploring very specifically what level of compe-tencies a proposed civil service commissionmight receive in a country, given the varying po-sitions of different stakeholders, and seeking toestablish a possible compromise point. However,an analysis that is exclusively focused at the mi-cro-level will clearly also have limitations, e.g. inrespect to understanding how the particular is-sue of concern is linked to wider country-leveldynamics.

As for levels one and two, PGPE for specificoperations can follow the basic framework setout in Part One and use the approaches dis-cussed in the annex. Here, we just point to a fewspecific issues to consider:



• For immediately project-focused GPE, theaim is most directly to have an output thatcan be used in operations, either in design oroperations management. In order to achievethis aim, it is important to integrate the per-son/team carrying out the political economyanalysis closely with the task team—to accessthe task team’s existing tacit knowledge, tounderstand operational concerns, etc. (seealso Part five above on process issues).

• Similar to sector-focused or thematic analy-sis, having existing macro (and relevant sec-tor) analysis is very useful to be able toembed the analysis in the ‘bigger picture’.However, even where such background ma-terial is missing, GPE analysis can still bevaluable for addressing issues of feasibilityand potential risks.

• This level of analysis may be combined withsocial development-types of analysis, in par-ticular analysis of the distributional impactsof reforms (Poverty and Social Impact Analy-sis or PSIAs). Some PSIA studies focused onspecific policy decisions and their likely dis-tributional impacts have included politicaleconomy analysis.19

• Also, at this level, it can be particularly use-ful to drill down by using the Senturion ex-pected utility stakeholder model (see Box 10in Annex 1).

II.4. SUMMARY: LEVELS OFANALYSIS

This section has spelled out how problem-drivenGPE analysis can be applied to different levels:the country level, the sector-focused or thematiclevel, and the project or individual policy level.In many cases, teams have undertaken GPEanalysis that straddles more than one of theselevels. This allows combining an analysis of thecountry-level ‘big picture’ with specific analysismore directly focused on questions linked toBank strategies and operations. This has for ex-ample been the route that the CGAC GPE pilotsin Zambia, Mongolia, and Mali have taken (as re-flected in the arrow on the right hand in Figure 3).

It is important to recognize that the nature ofthe benefit varies across levels. That is, the ben-efits of country-level analysis are different fromthose that can be derived from a sector or proj-ect-level use of problem-driven GPE analysis.Furthermore, each level presents somewhat dif-ferent challenges of doing GPE analysis. The in-tention of this section has been to draw atten-tion to the need to be aware of different levelsand hence different options that task teams maychose to pursue, to provide some broad guidanceand ideas, as well as to reflect some key existingexperiences for each level.

A N N E X 2


19. On Poverty and Social Impact Analysis, see: http://go.worldbank.org/39I9SFVEJ0.

Auty, R. 2007. Aid and Rent-Driven Growth—Mauritania, Kenya and Mozambique Com-pared. UNU-WIDER Research Paper No.2007/35. Available at: http://www.wider.unu.edu/stc/repec/pdfs/rp2007/rp2007-35.pdf.

Badkoubei, S. and K. Kaiser. 2009. “Gover-nance and Political Economy Issues inGrowth Analysis: A Survey of Recent WorldBank Country Economic Memoranda(CEMs).” Washington DC: World Bank(unpublished draft).

Basedau, M., and W. Lacher. 2006. “A Paradoxof Plenty? Rent Distribution and PoliticalStability in Oil States.” GIGA Working Pa-per No. 21. Hamburg: German Institute ofGlobal and Area Studies.

Besley, T. 2006. Principled Agents? The PoliticalEconomy of Good Government. Oxford: Ox-ford University Press.

Campos, J. E., and J. L. Syquia. 2006. “Man-aging the Politics of Reform: Overhaulingthe Legal Infrastructure of Public Procure-ment in the Philippines.” World BankWorking Paper No. 70. Washington, DC:World Bank.

Campos, J. E., and S. Pradhan (eds.) 2007. TheMany Faces of Corruption. Tracking Vulner-

abilities at the Sector Level. Washington,DC: World Bank.

Chabal, P., and J. P. Daloz. 1999. Africa Works:Disorder as Political Instrument. Blooming-ton: Indiana University Press.

Chibber, V. 2003. Locked in Place: State-Build-ing and Late Industrialization in India.Princeton: Princeton University Press.

Cox, G., and M. McCubbins. 2001. “The In-stitutional Determinants of Economic Pol-icy Outcomes.” In Presidents, Parliaments,and Policy, edited by S. Haggard and M.McCubbins (Cambridge: Cambridge Uni-versity Press).

de Mesquita, B., A. Smith Bueno, R. M. Siver-son, and J. D. Morrow. 2003. The Logic ofPolitical Survival. Cambridge, MA: MITPress.

Erdmann, G., and U. Engle. 2006. “Neopatri-monialism Revisited: Beyond a Catch-AllConcept.” GIGA Working Paper No 16.Hamburg: German Institute of Global andArea Studies.

Fritz, V. 2008. Applied Governance and GrowthDiagnostics: Basic Framework and Some Il-lustrations from Zambia. Washington DC:World Bank (draft mimeo).


References

Fukuyama, F. 2007. “Development and theLimits of Institutional Design.” In PoliticalInstitutions and Development, edited by V.Popov and N. Dinello. Cheltenham: Ed-ward Elgar.

Goodin, R. E., and C. Tilly (eds.). 2006. TheOxford Handbook of Contextual PoliticalAnalysis. Oxford: Oxford University Press.

Grindle, M. 2007. “Good Enough GovernanceRevisited.” Development Policy Review 25(5): 553–574.

Haber, S., A. Razo, and N. Maurer. 2003. ThePolitics of Property Rights: Political Instabil-ity, Credible Commitments, and EconomicGrowth in Mexico 1876–1929. Cambridge:Cambridge University Press.

Helmke, G., and S. Levitsky. 2004. “InformalInstitutions and Comparative Politics: AResearch Agenda.” Perspectives on Politics 2(4): 725–40.

IADB (Inter-American Development Bank)2006. The Politics of Policies. Economic andSocial Progress in Latin America: 2006 Re-port. Washington DC: IADB. Available at:http://www.iadb.org/res/ipes/2006/index.cfm.

IEG (Independent Evaluation Group). 2006.Annual Review of Development Effectiveness2006: Getting Results. Washington DC:WBG/IEG.

IEG. 2008. Public Sector Reform: What Worksand Why? An IEG Evaluation of World BankSupport. Washington DC: WBG/IEG.

Keefer, P., and S. Khemani. 2005. “Democracy,Public Expenditures, and the Poor: Under-standing Political Incentives for ProvidingPublic Services.” The World Bank ResearchObserver 20 (1): 1–27.

Keefer, P. 2007. “Beyond Legal Origin andChecks and Balances: Political Credibility,Citizen Information, and Financial SectorDevelopment.” Policy Research Working Pa-

per No. 4154. Washington DC: WorldBank.

Khan, M. 2006. Governance, Economic Growthand Development since the 1960s: Back-ground paper for World Economic and SocialSurvey 2006. New York: United Nations.Available at: http://www.un.org/esa/pol-icy/backgroundpapers/khan_back-ground_paper.pdf.

Khan, M., and K. Jomo. 2000. Rents, Rent-Seek-ing and Economic Development: Theory andEvidence in Asia. Cambridge: CambridgeUniversity Press.

King, G., R. Keohane, and S. Verba. 1994. “De-signing Social Inquiry.” Scientific Inferencein Qualitative Research. Princeton, NJ:Princeton University Press.

Krueger, A. 1974. “The Political Economy ofthe Rent-Seeking Society.” American Eco-nomic Review 64 (3): 291–303.

Lal, S. 2006. “Can Good Economics Ever BeGood Politics? Case Study of India’s PowerSector.” World Bank Working Paper No.83. Washington, DC: World Bank.

Lebanon Country Governance and Anti-Corrup-tion (CGAC) Process: Synthesis Report. 2009.Washington DC: World Bank (draft).

Leftwich, A. 2007. “From Drivers of Change tothe Politics of Development: Refining theAnalytical Framework to Understand thePolitics of Places Where We Work.” Notes ofGuidance for DfID Offices. University ofYork and DfID London (mimeo).

Molinas, J., Pérez-Liñán, A., Saiegh, S., Mon-tero, M. 2008. “Political Institutions, Pol-icymaking Processes, and Policy Outcomesin Paraguay.” In: Stein, E. and Tommassi,M. Policy Making in Latin America: HowPolitics shapes Policies. Washington DC:IADB and DRCLAS.

R E F E R E N C E S


Mongolia: The Political Economy of the ResourceParadox. 2009. Washington DC: WorldBank (unpublished).

North, D. 2007. A Conceptual Framework for In-terpreting Human History. http://cniss.wustl.edu/publications/2007north.pdf.

OECD. 2007. “Informal Institutions and De-velopment: How they Matter and WhatMakes them Change.” In: Informal Institu-tions: How Social Norms Help or Hinder De-velopment. OECD Development CentreStudies, No. 37. Paris: OECD.

OECD/DAC, 2009. Donor Approaches to Gov-ernance Assessments: Guiding Principles forEnhanced Impact, Usage and Harmonisa-tion. Paris: OECD. www.oecd.org/dac/gov-ernance/govassessment.

O’Neil, T. 2007. Neopatrimonialism and PublicSector Performance and Reform. London:ODI. http://www.odi.org.uk/PPPG/politics_and_governance/publications/GAPWP2BN1.pdf

Olson, M. 1965. The Logic of Collective Action:Public Goods and the Theory of Groups.Cambridge: Harvard University.

Ostrom, E. 1990. Governing the Commons: TheEvolution of Institutions for Collective Action.Cambridge: Cambridge University Press.

Ostrom, E. 2005. Understanding Institutional Di-versity. Princeton: Princeton University Press.

Pierson, P. 2004. Politics in Time: History, Insti-tutions, and Social Analysis. Princeton:Princeton University Press.

Ragin, C. 1987. The Comparative Method: Mov-ing Beyond Qualitative and QuantitativeStrategies. Berkeley: University of CaliforniaPress.

Rakner, L., L. Mukubvu, N. Ngwira, K.Smiddy, and A. Schneider. 2004. TheBudget as Theatre—The Formal and Infor-mal Institutional Making of the Budget

Process in Malawi. Bergen: ChristianMichelsen Institute.

Randall, V. 2007. “Political Parties and Demo-cratic Developmental States.” DevelopmentPolicy Review 25 (6): 633–652.

Rhodes, R. A. W, Sarah Binder, and Bert A.Rockman. 2006. The Oxford Handbook ofPolitical Institutions. Oxford: Oxford Uni-versity Press.

Rodrik, D. 2008a. “Spence Christens a NewWashington Consensus.” The Economists’Voice 5 (3). Published on the Web by TheBerkeley Electronic Press at http://www.be-press.com/ev/vol5/iss3/.

Rodrik, D. 2008b. “Second Best Institutions.”Unpublished paper available online athttp:// ksghome.harvard.edu/~drodrik/Sec-ond-best%20institutions%20paper.pdf.

Stein, E., and M. Tommasi. 2008. Policy Mak-ing in Latin America: How Politics ShapesPolicies. Washington and Cambridge: IADBand DRCLAS.

Tsebelis, G. 2002. Veto Players: How Political In-stitutions Work. Princeton: Princeton Uni-versity Press.

Weingast, B., and D. A. Wittman. 2006. TheOxford Handbook of Political Economy. Ox-ford: Oxford University Press.

World Bank, 2004. World Development Report2004: Making Services Work for Poor People.Washington DC: World Bank.

World Bank, 2007a. Strengthening World BankGroup Engagement on Governance and An-ticorruption. Washington DC: World Bank.http://www.worldbank.org/html/extdr/comments/governancefeedback/gacpaper-03212007.pdf.

World Bank. 2007b. Implementation Plan forStrengthening World Bank Group Engage-ment on Governance and Anticorruption.Washington, DC: World Bank.

R E F E R E N C E S


World Bank. 2007c. Tools for Institutional, Po-litical, and Social Analysis of Policy Reform:A Sourcebook for Development Practitioners.Washington DC: World Bank.

World Bank. 2008a. Benin: Constraints toGrowth and Potential for Diversification andInnovation—A country Economic Memoran-dum. Washington DC: World Bank.

World Bank. 2008b. The Political Economy ofPolicy Reform: Issues and Implications for

Policy Dialogue and Development Opera-tions. Washington DC: Social DevelopmentDepartment, World Bank.

World Bank. 2008c. Philippines—JointIBRD/IFC/MIGA Country Assistance Strat-egy FY2010–2012. Draft for upstream re-view.

Zambia: Governance, Political Economy and De-velopment Strategy. 2007. Washington DC:World Bank (unpublished document).

R E F E R E N C E S


Why is governance and political economy analysis important? Increasingly, it isrecognized that governance and political economy factors play a criticalrole in shaping development effectiveness and development outcomes.It is also important to gain better understanding of the environment—atcountry as well as at sector levels—in which World Bank operations are takingplace, to define feasible entry points and seize opportunities to promoteprogressive change.

Applying problem-driven governance and political economy analysis is not apanacea but should help with:

• Enhancing World Bank strategies (e.g. CASs, ISNs, sector strategies) andoperations, and improving their feasibility.

• Finding ways to change Bank programming if a ‘best practice’ technicalapproach has not been accepted, is not implemented and/or producesnegative unintended consequences.

• Supporting Bank staff to be more informed and effective interlocutors withclients.

• Speeding up the time taken to inform staff newly working on a country aboutrelevant governance and political economy contexts.

The objective of this good practice framework is to systematizeapproaches to governance and political economy analysis for WorldBank task team leaders and teams. As country and sector situations are verydifferent, this framework does not set out a particular product but insteadpresents a menu of options, proposes standards for diagnostics and synthesizeslessons learned.

The emphasis on a ‘problem-driven’ approach helps to make the analysisoperationally relevant, and comprises working through three layers: first,identifying the problem or opportunity to be addressed, second, mapping outthe institutional and governance arrangements and capacities and, third, drillingdown to the political economy drivers. This basic approach can be applied toanalysis at country, sector, or project levels.

Documents

Good Practice Framework - World Bank