1 
Final Report 
Performance Evaluation of 
PRODECENTRALIZATION II 
Client: USAID 
November 2012 
2 
Index
 Página 
Executive Summary 03 
Introduction 13
Chapter 1: Peruvian Context 15
Chapter 2: The national focus: From policy to practice 20 
 
2.1 The regulatory framework 20 
 2.2.1. Objectives of the project 20
 2.1.2. The program’s contribution on legislative matters 20 
 2.1.3. Aspects that limited or facilitated the program’s action on that 
issue 23 
 
2.2 Social Conflicts 24 
 2.2.1. Objectives of the project 24
 2.2.2. The program’s contribution on the issue of social conflicts 25 
 2.2.3. Aspects limiting or facilitating the program’s action on this 
point 27 
2.3 The contribution from a gender perspective 28 
Chapter 3: Capacity Building 30
3.1 Objectives and indicators of the capacity-building program 30 
3.2 Quantitative research 30
3.3 Qualitative evaluation of capacity building 41
 3.4 Main findings 45
Chapter4 Evaluation of the program characteristics 47
Chapter 5 Conclusions 53
Chapter 6 Lessons learned 56 
Appendix 
Theoretical framework of the study 59 
3 
 Executive Summary
The USAID/Peru ProDecentralization (PRODES) program supports the efforts of the 
Peruvian government in the decentralization process through the improvement of public 
policy and the regulatory framework, as well as strengthening the capacity of sub￾national governments. 
The purpose of this study is to assess the performance of ProDecentralization II during 
the period 2008 and 2012. The evaluation objectives were as follows: 
 Assess the effectiveness of ProDecentralization results to date at the national 
and sub-national (regional and local) levels. 
 Identify factors that facilitate or create challenges for the achievement of 
ProDecentralization objectives. 
 Identify lessons learned and make strategic recommendations for how USAID can 
better provide technical assistance to the GOP to strengthen the decentralization 
process. 
The study considered the following evaluation questions: 
 To what extent (quantitative and qualitative) has ProDecentralization II achieved 
overall results/outcomes with respect to: 1) improving the decentralization 
regulatory and policy framework; and 2) strengthening sub-national governments’ 
capacity to govern? 
 What were the primary obstacles (internal to the project and external) to 
achieving ProDecentralization’s expected outcomes? 
 Which are the main lessons learned and recommendations for USAID and the GOP 
for possible future interventions to improve decentralization? 
The study was carried out in the departments of Ucayali, San Martin and Ayacucho. In 
these regions, we focused on doing an assessment at the regional, provincial and district 
levels. 
With this framework in mind, we designed an assessment instrument that includes a 
qualitative component (in depth interviews) and quantitative research based on 
secondary data. In the case of the quantitative assessment, we employed a panel data 
approach. With USAID’s consent, the Assessment Team proposed that quantitative 
assessment be centered on two indicators (Y variables) to study the performance of the 
ProDecentralization Program in the regional governments of Ayacucho, San Martin, and 
Ucayali and in each of their respective provincial municipalities. The indicators were as 
follows: 
4 
a. Efficient investment budget execution; and 
b. Number of public investment projects approved. 
In addition, a qualitative study was conducted to identify limitations, bottlenecks and 
opportunities to achieving ProDecentralization’s objectives. Interviews were conducted 
at each level of government: national (29 interviews), regional (10 interviews per 
region) and local (8 interviews per region). The guidelines applied to these interviews 
can be found in the appendixes. 
The evaluation results are grouped below by evaluation question. 
ProDecentralization II Results 
Findings 
Regulatory Framework and Public Policy 
 ProDecentralization’s support on regulatory framework and public policy matters 
has been necessary– in the perception of the stakeholders themselves- but 
insufficient to promote key reforms. Political will on the part ofdecision makers 
is also necessary improve the decentralization process. 
 Establishing a working relationship with the Assembly of Regional Presidents was 
a positive aspect of ProDecentralization that will help drive further progress on 
regulatory and policy issues. 
 The Decentralization Compass is valued by stakeholders because it has opened up 
a space for interaction and coordination that was previously lacking. 
 While the quality of the Decentralization Compass was consistently highlighted by 
respondents, it is noteworthy that it failed to generate a consensus for an agenda 
to strengthen the decentralization process. Reaching this type of consensus is 
necessary to influence legislation. 
 The documents or contributions that were submitted to implement regulations, 
such as the Unified Administrative Procedures Text (TUPA) model and regulations 
governing associations were not mentioned by interviewed stakeholders. High 
turnover of personnel reduced the visibility and level of awareness of these 
contributions. 
 The Decentralization Map, intended to provide the capacity to monitor the 
decentralization process, was not mentioned as a reference document that 
clarifies policy reform for the Executive and Legislative Branches of the Peruvian 
Government. 
 Technical assistance to develop legislation related to the decentralization 
process was valued by stakeholders. 
 The assistance provided to the regional governments received greater recognition 
than that given to provincial and district municipalities; these more local 
5 
governments wanted greater input during discussions on the decentralization 
process. 
Social Conflicts 
 While national guidelines for conflict prevention have been developed and the 
organizational structure of the relevant government offices has been 
strengthened, these advances do not resolve fundamental underlying problems. 
 In the absence of a shared vision about how to achieve a system that responds 
well to dealing with conflicts, assessments by stakeholders of accomplishments to 
date and demands for future action are highly diverse. 
 The establishment of an office devoted to managing social conflicts should be 
supplemented by an effective reorganization, driven by a deeper analysis of 
conflict situations, of every relevant level of government. 
 The project is linked to capacity building in seven regionsassociated with the 
handling of social conflicts. 
Gender 
 The gender approach achieved positive results on personal and institutional (i.e. 
programs oriented at gender gap reduction) levels and also increased awareness 
of women’s and children’s issues. 
 Fourteen projects prepared by recipients of ProDecentralization’s public 
investment projects diploma included a gender approach (4 done in Ucayali, 3 in 
Ayacucho and 4 in San Martin), for example, the “Sexual and Reproductive Health 
for teenagers Capacity Strengthening in the districts of Callería, Yarinacochas 
and Mamantay, Coronel Portillo Province in Ucayali.” 
 Fifteen draft ordinances used a gender approach, including issues such as 
maternal health, children, and teenagers. One from Maynas Province, District of 
San Juan Bautista, referred to the strengthening of the services to raise 
awareness and prevent domestic violence. As of June 2012, 01 was approved in 
this report’s scope of evaluation1
, which declares as a priority and of public need 
the attention of early childhood in Apurimac. 
Capacity Building 
 The PRODES proposal aligned with the needs and priorities of regional and 
municipal governments. 
 Training had better results in places where it was accompanied by technical 
assistance. 
 
1
On August 20th, 2011, the Regional Council of Ayacucho, approved Regional Ordinance No. 021-
2011-GRA/cr under which the improvement of neonatal maternal health is declared public policy. 
6 
 Training on the topic of TUPA was judged to be more effective to the extent that 
it was treated as a technical subject and directly associated with revenue 
generation. 
 Respondents were less positive about training on the topics of budgets, contract 
hiring and procurement. It was pointed out that political forces often prevented 
projects, hiring and procurement processes from being carried out within the 
established regulatory framework. The problem is not limited to the lack of 
familiarity with the regulations or procedures, but extends to political pressure 
to act contrary to these same regulations or procedures. 
 Those interviewed at the sub national levels highlighted the quality of the 
training sessions, the methodology, the teachers selected and the designed 
materials. 
 In addition to the significant number of government staff and officials that 
received training, the materials and methods have also been incorporated by 
other institutions working within the country and, consequently, have been 
reaching more beneficiaries. 
 This evaluation shows empirically that ProDecentralization increased the number 
of local projects approved by the GOP project approval process. 
 Training local elected officials (such as city council members) appears to have 
had a positive impact on their productivity since those that were trained are 
performing their job functions better than their predecessors. 
 The quantitative analysis indicates that the program has had a positive impact on 
the number of public investment projects that have been approved. According to 
the qualitative data gathered, this effect is due to innovative training 
supplemented by good materials. All those interviewed recognize the positive 
attributes of the training. 
 With respect to the effects on the indicator “execution of the investment 
budget”, the quantitative study failed to provide conclusive results. One possible 
explanation for this may be that the impact of investment budget execution is 
harder to measure, since it does not rely on the governments’ capacities alone. 
Moreover, this variable does not include other actions taken by PRODES or other 
projects sponsored by USAID, the MEF and various other donors who support the 
strengthening of public administration in this same scope of reference. 
Characteristics of the PRODES II Projects: 
 The demand-based approach, allowing the stakeholder demand to drive 
activities, improved versatility and allowed PRODES to address social conflicts, an 
issue that was not initially included in the project. However, the demand based 
approach made it difficult to measure outcomes and led to a dispersion of 
efforts. 
7 
 The gender based approach that has been part of the training and activities 
implemented by PRODES has encouraged actions aimed at reducing the gender 
gap or are otherwise beneficial to women. 
 The gender approach has enabled female elected local officials to identify and 
take action on gender issues. 
 Building alliances has improved achievement of objectives and optimization of 
resources. 
 There was no overlap with other foreign assistance agencies or other North 
American foreign assistance projects. 
 The perception is that the project is completely aligned with the characteristic 
that is often attributed to North American foreign assistance: The interest in 
strengthening the governing bodies of the nation. 
Conclusions
Regulatory Framework 
 The decentralization process is currently mired in stagnation and inactivity due 
to a lack of focused direction from the Central Government. This situation will 
affect the execution of the next phase of PRODES as well as expectations for the 
Project. 
 The progress toward decentralization that has occurred over the past few years 
has been largely driven by regional governments. Their role as catalysts of this 
process could be supported and encouraged with technical assistance. 
 In order to ensure that the decentralization process moves forward on legislation, 
political spokespersons are needed. It is essential that regional and local 
governments possess a greater capacity to develop proposals and take a 
leadership role. 
 The diagnosis performed using the Decentralization Map and the Decentralization 
Compass is still valid which would indicate that the data collection process was 
adequate. 
 The technical assistance used to prepare legislation has proved to be effective in 
overcoming the limitations faced by international foreign assistance when it 
wants to influence the legal framework. Resistance from multiple sources is 
reduced using this approach. 
 Fostering the creation of spaces for dialogue between the different levels of 
government was one of PRODES’ most effective interventions. 
 Lack of clarity as to the responsibilities of the different levels of government 
generates conflict, which has an impact both on the management of the 
decentralization process and on conflict resolution. 
 Political dynamics in the decentralization process carry a good deal of weight. 
Because of this, an approach which is exclusively technical in nature, with a 
8 
central approach aimed at building capacities is insufficient. This is particularly 
clear when dealing with capacity building or creating a structure to resolve social 
conflicts. 
 While the submission of proposals for ordinances that involve gender has 
increased, the rate of approval was significantly lower for gender related 
ordinances (of the 15 ordinances submitted; only 2 were eventually approved as 
of June 2012). 
Sub-national Government Capacity building 
 The intervention by PRODES in the presentation of projects is seen as positive. 
We have also identified government administrative actions that demonstrate a 
notably greater level of efficiency as a result of activities implemented by 
PRODES. 
 Capacity building to improve government efficiency requires greater emphasis on 
technical assistance. Technical assistance increases the likelihood that individual 
knowledge translates to institutional change. 
 The strengths of the PRODES approach includes the use of alliances, its 
development of innovative methods, designing quality materials and meeting the 
needs identified and defined by the beneficiaries. 
 The weaknesses of the PRODES model are 1) the heterogeneity of interventions 
which makes it difficult to measure impact, 2) The governments who expressed 
their demands have not always possessed much clarity about their needs, 3) 
There were difficulties to achieving the participation of the highest level public 
officers in trainings, and 4) the high rate of staff turnover at the levels of 
government where PRODES worked. 
 Capacity building is still in demand at different levels of government. 
 Encouraging the participation of women in capacity building courses was an 
appropriate measure given that women usually have a lower level of education 
and experience than their peers. 
Characteristics of the PRODES II Projects 
 The demand based approach is a unique aspect of PRODES. However, there is no 
evidence to suggest that the program’s achievements owe themselves to this 
aspect. 
 The demand based approach succeeded in winning the trust and commitment of 
the beneficiaries. 
 The demand based approach does not have comparable effects at the various 
starting points of each zone. The quantitative and qualitative data reflect 
differences from one intervention location to another. 
 The demand based approach requires a complimentary in-depth institutional 
analysis to improve the effectiveness of the interventions. 
9 
 The work done with many regional governments, provincial municipalities and 
district municipalities created valuable spaces for intergovernmental dialogue. 
However, the resulting dispersion of activities and resources reduced the direct 
impact of the activities. 
 PRODES has carried out activities in areas where there are other USAID sponsored 
programs and where other international foreign assistance agencies operate, 
without incurring any overlap. There was a case with a North American foreign 
assistance agency where efforts were coordinated to complement each other: 
ProParticipation (aimed at the population) and ProDecentralization (aimed at the 
authorities). 
 PRODES strengthens the North American foreign assistance approach, which is 
strongly concerned with institutional strengthening. 
Bottlenecks 
 The problems that the Project deals with are technical in nature; the 
decentralization process is political and creates discrepancies between different 
levels of government, particularly regarding issues related to transfers of power 
and resources. 
 The lack of consensus prevails across different levels of government. There is no 
consensus about what needs to be done on the topics of decentralization, 
handling social conflicts or administrative systems like SNIP. This lack of a shared 
vision produces both progress and setbacks which reduces the likelihood of 
obtaining results. 
 High staff turnover is detrimental since it impedes the accumulation of 
knowledge and experience. In this regard, the training provided loses impact and 
sustainability. 
 The current level of political instability in the regions where PRODES worked 
limited the probability of the success of these interventions. 
 Finally, one of the principal limiting factors to the decentralization process has 
been the absence of mechanisms for intergovernmental coordination, the 
difficulties in forming consensus and the lack of leadership and political vision on 
the part of the executive and other key stakeholders. The project’s role in 
creating spaces for intergovernmental dialogue is duly noted. 
Lessons Learned: 
 It is essential that we maintain and even accentuate the interview process 
carried out by PRODES (up-down and bottom-up). This type of process, which 
received high marks from the stakeholders, becomes necessary when seeking 
greater participation on the part of the regional, provincial and district levels of 
government in determining the regulatory framework. 
 A project’s resources can be optimized by using an approach of alliance building. 
10 
 Particular attention must continue to be paid to the selection of trainers and 
getting government institutions involved in the preparation of materials. 
 Projects such as PRODES should place more emphasis on technical assistance. It 
improves results and gives the work a long term perspective. 
 The timing of changes in government administrations must be taken into account 
when designing the programs. Training programs that are started during this 
phase weaken the program intervention’s overall impact due to inevitable staff 
turnover and because complications often arise during the transition period. 
 Having a tailor made program to offer (in this case a demand based approach) 
serves to build bonds of trust with the different bodies of government with whom 
the project will work and makes it easier to adapt to the situation. But a prior 
assessment of the government bodies that will be participating in this program is 
essential in order to determine if they have the capacity to clearly identify their 
needs. 
 We recommend that the gender based approach be maintained and even given 
more emphasis since this ensures that gender issues are given appropriate 
consideration. 
Recommendations 
 We recommend to continue providing neutral spaces for dialogue between key 
players in the National Government (Executive and Legislative Branches) with a 
view toward forming networks and developing synergies that carry the potential 
to build consensus about the direction of the decentralization process and the 
direction of related legislation. 
 It is important to continue to develop consensus and coordinate existing 
proposals about guidelines for the decentralization process at the sub-national 
levels of government. This must be done in a way that promotes a Bottom–up 
dynamic to form coalitions for policies that move the process forward. Along 
these lines, the Alliance with the Assembly of Regional Governments and 
associations of municipal government should be strengthened. 
 The Decentralization Compass which was recently transferred to the 
Congressional Committee on Decentralization should be monitored to ensure its 
continuity. 
 More emphasis should be given to technical assistance focusing efforts at the 
regional level. 
 It is recommended that PRODES continues its work and even increase its 
assistance to initiatives such as the association of townships and regional 
alliances. 
Social Conflicts
a. Capacity building must consider the unique characteristics of each place 
where we implement our program. The majority of these conflicts are 
11 
associated with a particular scope of influence such as the existence of 
mines, forests, etc. 
b. Capacity building should proceed when directly related to the powers 
conferred to the sub-national governments. Doing it before these powers and 
authority are clearly defined would be ineffective. 
c. Indicators are needed to enable the measurement of the capacity of sub￾national governments in conflict prevention and management. 
d. While it is true that the most obvious social conflicts are associated with 
mining operations and large scale infrastructure projects, there are a number 
of totally different conflicts which cannot be ignored because doing so would 
increase unrest in populations that have been neglected. 
e. It is necessary to examine ways to strengthen political representation since its 
absence is often a cause of the emergence of unrest and conflict. 
f. More coordination needs to be encouraged between the National Government 
and sub-national governments on matters of conflict resolution. Moreover, 
greater clarity is needed about the definition of the powers and jurisdiction 
held by each body and level of government. 
 With regard to the gender issue, we recommend that the approach be enhanced 
since the presentation of gender related initiatives by itself does not guarantee 
subsequent approval. This requires the formation of alliances to exert the 
necessary political and social pressure to achieve their approval. 
Recommendations to USAID: 
 With regard to legislative reforms that require political willingness and a high 
level consensus, we recommend that USAID search for other types of alliances 
that effectively enable the key players to incorporate their proposals and vision 
into the debate. It must be taken into account that regional governments lack 
the resources and access to experts to be able to submit proposals that modify 
the current regulatory framework. 
 One relevant challenge consists of improving the availability of data required to 
conduct a rigorous assessment prior to program implementation. In this sense, 
the barometer should already have information in its data base about the number 
of PIPs (Public Investment Projects) that have been prepared in order to establish 
the ratio, N° of Investment Projects approved/prepared, which is a better 
indicator of efficiency. 
 We suggest that certain aspects of the demand based approach be revisited. 
Beyond its inherent advantages, the approach is only effective where 
governments correctly identify their needs. In this sense, the demand based 
approach could be supplemented with an assessment to ensure that training 
content is adapted to the needs of the regional and municipal governments and 
not solely based on an open ended request for training. This point becomes more 
relevant in the case of a variable which is difficult to model such as “investment 
budget execution.” 
12 
 In the future, program design, implementation, and data collection should be 
conducted with a view toward performing a rigorous evaluation. Data collection 
should be consistent with the evaluation method proposed and defined from the 
beginning of the program. In this regard: 1) we suggest reducing the 
heterogeneity of the training, and 2) keeping more detailed and computerized 
records on the profile of each trainee and of participating government entities. 
 Recommendations to the Peruvian government 
 The Government should recognize that, in practice, the decentralization process 
has been moving forward with different degrees of efficiency between sectors 
and regions. This creates the need to have a clear diagnosis of the project’s 
status and an investment that should be made in design and monitoring. 
 It is essential to clearly convey the direction that the decentralization process 
might be headed or where stakeholders want it to go. This involves clearly 
defining the functions, powers and mechanisms of interaction between the 
different levels of government. 
 Encourage Regional Government Presidents to assume a leadership role in their 
areas of jurisdiction (provincial and district municipalities). 
 Place more emphasis on technical assistance in the regions and municipal 
governments; initiatives such as those in CONECTAMEF can contribute to this 
objective. 
 It is essential to create new spaces where people are able to interact and 
coordinate with the different government bodies. 
 Look at developing legal frameworks that can deal with the issues behind social 
conflicts. Understand that the Central Government also needs to adapt its 
procedures and training to learn how to operate in the scenario of a 
decentralized nation. 
 Place more emphasis on developing indicators that enable all stakeholders to 
more clearly read the status of the decentralization process. 
13 
Introduction
Since its inception in 2002, the decentralization process has been a fundamental aspect 
of the reform of the Peruvian state and participatory democracy, with emphasis placed 
on it during the administrations of Alejandro Toledo and Alan Garcia. In the last ten 
years, significant progress has been made on the decentralization process with the 
support of different levels of government and civil society. 
The main goal of the Prodescentralization project was to support efforts to go forward 
with the decentralization process, focusing on two main results: 
1) Improvement in the legal and public policy framework (national level) 
2) Building local governments’ capacity to govern effectively (sub-national level) 
Activities at the national level were designed to improve the legal and public policy 
framework for decentralization. The main activities were the following: 
1) Evaluation of the legal and public policy framework 
2) Design of a tool for prioritizing actions related to the legal and public policy 
framework 
3) An annual status report on decentralization 
4) Technical assistance on critical legal reform matters and the implementation of 
regulations, procedures and public policies 
At the sub-national level, the focus was on building local and regional governments’ 
capacity to govern effectively, with the following as the main intervention areas: 
1) Reinforcement of Good Government Practices such as participatory budgeting 
and planning, including public accountability processes. 
2) Improvement in the implementation of National Administrative Systems. Efforts 
were focused on strengthening the National Public Investment System, 
procurement processes, state contracting and results-based budgeting. 
3) Improvement in the efficiency of selected municipal governments, including 
administrative simplification and results-based budgeting. 
4) Promotion of local economic development. 
5) Implementation of innovative strategies for tax collection at the local level. 
6) Support for municipal associations. 
7) Support for Inter-regional Coordination Boards. 
8) Formation of public / private alliances. 
The project used the following implementation strategies: 
 Demand-based project: Taking into account the priorities and needs stated by 
national public policy makers and regional and provincial authorities. At the 
beginning of the project, an assessment of regional and provincial governments’ 
needs and priorities was prepared. Based on this information, the project was 
14 
developed. During the second year, an instrument called a “Governance 
Barometer” was designed for following up on "measurable" improvements, along 
with other monitoring and evaluation tools. The assessment was updated 
annually to ensure that the technical assistance met Peruvian government 
entities’ needs and priorities. 
 From policy to practice: While the activities were organized under the 
aforementioned headings, every activity executed by Prodescentralization had 
applied components; that is, the project provided support on the development of 
national policies and/or procedures for implementing these policies, and sought 
their application in practice with selected regional and provincial governments. 
 Alignment with Peruvian government policy and promotion of ownership: Every 
initiative implemented by the program (except for the annual decentralization 
analysis), was executed in collaboration with the responsible Peruvian 
government entities according to the subject matter or intervention area. All 
training materials were approved and “co-branded” with the corresponding 
supervisory entity. 
 Results-based approach: sub-national interventions sought improvement in public 
administration. Regional and municipal governments demonstrated their 
commitment by applying the knowledge, techniques, tools and systems 
developed during the project. 
 Alliances and agreements were developed to maximize the impact of the project. 
The list of actors that worked with the project includes other USAID projects, 
other donors, civil society organizations, Peruvian government entities, 
associations at the regional and municipal levels and private entities. 
The information presented below analyzes all of the aspects detailed above, in an effort 
to establish the extent to which the strategies established initially were maintained and 
how this contributed (or not) to the achievement of the established objectives. Chapter 
1 is a contextualization of the decentralization process, which frames the activities 
carried out by the project. The second chapter covers the work performed in relation 
with the regulatory framework, while the third chapter presents an analysis of what has 
been achieved in terms of capacity building. The last chapter analyzes general 
characteristics of the program. 
The performance evaluation of this program was performed by a four-member team 
whose members were from Imasen and the Center for Social, Economic, Political and 
Anthropological Research?(CISEPA). The team, whose members were Edmundo Beteta 
(Team Leader), Giovanna Peñaflor (Qualitative Analyst), Jessica Benza and Cristian 
Alvan (Senior Analysts), reviewed the materials developed by the project, gathered 
secondary information, conducted interviews and analyzed the foregoing in order to 
prepare this document. 
 
15 
The Peruvian context
The decentralization process in Peru began with the Reform Act on Chapter XIV of Title 
IV of the 1993 Constitution (Law 27680) and the passage of the Framework Law on 
Decentralization (Law 27783) (LBD) in 2002. Within this framework, the decentralization 
policy was defined as a top-down process that would take place through an orderly, 
progressive transfer of powers and functions of the national government to regional and 
local governments in phases. The purpose of the process was to "achieve harmonious and 
sustainable integral development of the country through the separation of powers and 
functions, and the balanced exercise of power by the three levels of government for 
the benefit of the population" (LBD: art. 3). 
The basic regulatory framework (Constitutional Reform Law, Framework Law on 
Decentralization, Organic Law on Regional Governments, Organic Law on Municipalities, 
Law on the National Accreditation System and its Regulations, Law and Regulations on 
Incentives for the Formation of Regions, and the Fiscal Decentralization Law). During 
this period, investment and social programs were transferred to regional and local 
governments. 
In 2005 there was an unsuccessful attempt to form regions. Then, starting with the 
change of administration in 2006, there was a “decentralization shock" in which 
transfers of functions and assets, including the regional directorates of ministries, were 
accelerated and 95% of the transfers were completed. During this period, efforts were 
made to optimize the decentralized operation of administrative systems, particularly the 
National Public Investment System. The National Council for Decentralization, which was 
the joint technical body (with the involvement of the national government and regional 
governments) in charge of coordinating and supervising the decentralization process, 
was disbanded and replaced by the Decentralization Secretariat within the Office of the 
President of the Council of Ministers (PCM), as well as the newly created 
Intergovernmental Coordination Council (ICC) as an advisory body, with the participation 
of representatives of regional and local governments. 
From 2007 to 2008, progress was made on the passage of the Organic Law on the 
Executive Branch (LOPE) and the issuance of 100 legislative decrees 2 within the 
framework of the approval of the Free Trade Agreement with the United States. Among 
these, the following merit mention: Civil Service reform, the optional modernization of 
regional and local governments, and national strategic planning. (Prodes: 2008, p. 12). 
In addition, the implementation of the Results-Based Budget began and Congress passed 
new legislation to increase the development of municipal associations and to establish 
 
2
 Among them D.L. No. 1023, which created the National Civil Service Authority, the supervisory body for 
the Administrative System for Human Resource Management; D.L. No. 1024, which created and regulated 
the Public Manager Corps; D.L. No. 025, which approved training and performance regulations for the public 
sector; D.L. No. 1026, which established a special optional system for regional and local governments 
wishing to implement integral institutional modernization processes; and D.L. No. 1088, Law on the National 
Strategic Planning System and the National Center for Strategic Planning (PRODES:2008). 
Chapter 1
16 
the division of powers between the regional executive and legislative branches (or 
Regional Council) within Regional Governments. It also passed the Law on Equal 
Opportunities for Men and Women, which establishes guidelines for the different levels 
of government including regional and local governments, in order to ensure the principle 
of equality within the framework of their authority (PRODES: 2009, p. 10). 
In 2009, a debate and review of the proposed Law on the Organization and Functions of 
Ministries of the Executive Branch, which details shared and exclusive powers and 
functions, began. Through Law 29341, municipal associations were granted legal status, 
allowing them to execute projects and resources jointly. Similarly, Law 29379 
reinforced the Interregional Coordination Boards (JCI) as a mechanism to ensure 
continued progress on the regionalization process, which had stalled in 2005. 
As the decentralization process had not concluded, the qualitative research sought to 
delve deeper into the actors’ perception of the decentralization process. For this 
reason, questions were included to determine their opinion on the progress made and 
limitations of the process, as well as the main bottlenecks encountered. The following is 
a summary of some of the prevailing opinions of key actors on the context and current 
status of the decentralization process, since we consider them to be an important 
element to take into account in framing the intervention of PRODES-II, as well in 
reviewing assumptions for future intervention: 
 A recurring theme mentioned by respondents was the lack of proper distribution 
of powers and responsibilities among the different levels of government. 
Although decentralization was conceived at its outset as an orderly transfer 
process based on capacity-building in sub-national governments, respondents 
believe that, in practice, the process may have been influenced by political 
negotiation among the various actors, which resulted in a rather disorderly 
process in which powers have not been well defined. Although the different 
levels of government coexist peacefully, the existence of overlapping functions 
that hinder policy development and prevent proper use of public resources is still 
mentioned. 
 As the Framework Law on Decentralization (LBD) (Law 27783 of 2002) does not 
precisely define exclusive, delegated and shared powers or delineate functions, 
these powers have been defined by a multitude of pieces of legislation passed 
throughout the process, such as the Organic Law on Regional Governments 
(LOGR) (Law 27867 of 2002), the Organic Law on Municipalities (LOM) (Law 27972 
of 2003), the Organic Law on the Executive Branch (LOPE ) (Law 29158 of 2007) 
and even different laws on the organization and functions of ministries. 
Respondents agree that, in practice, timing differences in the passage and/or 
amendment of these laws has resulted in a disorderly process. According to 
functionaries of the Commission for the Decentralization of Congress and other 
experts, the fact that the laws on the organization and functions of ministries 
determine and define a set of functions shared by the three levels of government 
has given rise to sectoral logic in the process that is inconsistent as a whole. 
 If the main objective of decentralization is to "achieve integral and harmonious 
development of the territory through proper allocation of powers for the benefit 
of the population," respondents coincide that one of the main difficulties in 
17 
moving the process forward is precisely the lack of a clear division of powers and 
the difficulty of dealing with the problem of the diversity of sub-national 
governments, especially municipalities, with technical criteria when defining 
powers, functions and the allocation of resources. 
 The lack of clarity on the matter of powers limits citizens’ possibilities of 
identifying which level of government is responsible for policies, thereby 
hindering citizen oversight and the ability to hold authorities accountable. 
 Regarding the regulatory framework for sub-national governments, respondents 
agreed on the need to determine the role of the Regional Coordination Councils, 
as they did not have a clear understanding of their role with regard to 
coordination with mayors and the participation of civil society. 
 In addition, several respondents indicated that the organizational structure and 
institutional design of sub-national governments should be revised to ensure that 
they are able to operate effectively. 
 Another problem hindering the decentralization process is stagnation on the 
matter of regionalization and difficulties in creating macro-regions. However, 
rather than a regulatory problem, they consider this situation to be a political 
problem and they mentioned the correct decision by PRODES-II to contribute to 
the ongoing articulation process through the Interregional Coordination Boards 
(JCI) and to support the creation of municipal associations. 
 Respondents repeatedly mentioned that the stagnation of the regionalization 
process has had an impact on the lack of progress on the fiscal decentralization 
process accompanying the transfers, as it was originally tied to the 
regionalization process that stalled in 2005. The passage of a new law on fiscal 
decentralization is a pending issue that depends on political will on the part of 
actors in the Peruvian government. 
 Respondents also noted the absence of intergovernmental coordination 
mechanisms, the difficulty of reaching consensus and the lack of leadership and 
political foresight on the part of the Executive Branch and other key actors as a 
major constraint in the execution of the decentralization process. 
Decentralization has resulted in a proliferation of politically empowered actors, 
but the main problem is to answer the following question: How can a 
decentralized state with three levels of government be built? In the opinion of 
some respondents, it will be a long-term process that could take up to thirty 
years and will require a shared strategic vision of the process and the type of 
country and regions or municipal governments we want to have. 
 Regarding supervision of the process, respondents agree on the weakness of the 
former National Council for Decentralization (CND) and its failure to promote 
regionalization. However, they also indicate the limitations of the current 
Decentralization Secretariat (SD-PCM) created to replace it, as it is a third-level 
body with difficulties negotiating on equal terms with the ministries. Regarding 
the Intergovernmental Coordination Council (ICC), they mentioned that it was not 
18 
working; consequently, bilateral relations between the Executive Branch and 
regional/local actors were given priority. 
 These difficulties in intergovernmental coordination are repeated at the regional 
level, where respondents indicated that there is no capacity to articulate the 
relations of provincial and district governments. The ill-defined role of the 
Regional Coordination Councils as spaces that confuse intergovernmental 
coordination functions with those of the participation entity also hampers this 
process. 
 In addition, respondents indicated that there are also difficulties in intersectoral 
coordination within the Executive Branch, which hinders the coordinated 
execution of the decentralization process. This problem takes on additional 
importance when we consider that many of the functions of the powers shared by 
all three levels of government are determined by the laws on the organization 
and functions of each ministry. 
 Another element of the context that was mentioned in the interviews and forms 
part of the assessment on which foreign assistance interventions are conditioned 
is the lack of guidelines and basic consensus among the main actors. In this case, 
the need to coordinate and articulate proposals made by regional governments, 
as well as to give more prominence to the National Assembly of Regional 
Governments as the space for reaching consensus at the sub-national level and as 
a link with the Executive Branch at the national level, was also mentioned. 
The elements characterizing the abovementioned decentralization process are 
summarized in the following chart: 
19 
The national focus: From policy to practice
In this chapter of the report, matters related to the work of PRODES II regarding 
legislation and the implementation of polices, what it has done in relation with social 
conflicts and its contribution with regard to gender are analyzed. 
2.1 The regulatory framework 
2.1.1 Objectives of the project 
The intervention of PRODES-II in the development of the regulatory framework for 
decentralization and policy implementation has been defined with a "from policy to 
practice" approach, through which it seeks to collaborate with key actors in the Peruvian 
government (GOP), such as the Executive Branch (PCM, MEF and other sectors) and the 
Congressional Commission on Decentralization, to facilitate the development of 
regulations and the implementation of policies (policy) that contribute to improved 
performance of regional and local governments (practice) and ultimately, to progress on 
the decentralization process. 
Consequently, with regard to the regulatory framework for decentralization, the 
objectives of PRODES-II were geared toward: 
a. Providing information to key decision makers in the central government (GOP), 
thereby facilitating informed decision-making. 
b. Helping to build consensus among key decision makers in the Peruvian 
government (GOP). 
c. Promoting public debate on decentralization policies in order to identify needs 
for regulatory reform and to reach consensus on the general outlines of the 
decentralization 
2.1.2 The program’s contribution on legislative matters 
As shown in the PMP table (see Annex) in the preceding section, the project produced 
the annual “Decentralization Compass”, the Map of Legislation and Policies on 
Decentralization and an annual report on the decentralization process. 
The "Decentralization Compass" was developed in collaboration with key actors in the 
Executive and Legislative branches, regional and local governments and other relevant 
actors. In it, policies that should be promoted in order to achieve progress on the 
decentralization process were identified, providing information to policy makers that 
will improve the quality of the policy formulation process and build consensus. For 
example, for the last year of the project, with the help of the key actors mentioned 
above, the compass identified the following policies that should be promoted: 
Chapter 2
20 
a. Political articulation and coordination in the management of various government 
agencies; that is, alignment of policies, guidelines, plans and actions at both the 
national and decentralized levels. 
 
b. Define the steering role of the National Government and its implementation and 
operational mechanisms, taking into account the autonomy of decentralized 
governments. 
c. Decrease institutional parallelism. It would be necessary not only to transfer 
functions and powers from the National Government, but also to avert the 
emergence of parallel institutionalism which would give rise to the execution of 
national programs that overlap the programs run by decentralized governments. 
d. Incorporation of the notion that decentralization not only involves the reform of 
the organizational structure of the Executive Branch, but also its modus 
operandi, decision making and management systems, among other reforms. 
It should be noted that the comparison between these proposed topics and the answers 
obtained from the different actors interviewed for this report confirms the need to 
implement them. Regarding the need to incorporate changes in the Executive branch’s 
way of operating, positions such as that of the director of an NGO were found: "I think 
there are weak points in the regulatory framework, which are the laws that should 
apply to the ministries, the laws on the organization of the ministries’ functions, 
which should apply starting in the year 2007 when the Organic Law on the 
Executive Branch was passed. It states, ‘From now on, every ministry must be 
restructured, with new laws in line with decentralization.’ That took place in some 
ministries, but not in others. Therefore, there are ministries that still operate 
according to laws passed prior to decentralization," which indicates the need for the 
central government to review what they are and are not doing. In addition, the lack of 
leadership of the process on the part of the Executive branch was mentioned: "The 
central government has not taken on the task of strategic leadership in achieving 
articulation of the three levels of government. Articulation not only implies the 
operation of the Intergovernmental Coordination Council; it also means that in a 
specific territory, the three levels of government must be articulated, and the 
national government should take the initiative." (Congressional functionary) Although 
his remarks place greater emphasis on the operational aspect than the regulatory 
aspect, they show that he perceives the lack of political articulation and coordination, 
which is a problem related to the disappearance of the CND, as it left a vacuum that was 
not filled by the creation of the Decentralization Secretariat. This last situation leads 
some to think that the model used needs to be completely overhauled ("We already had 
a faulty framework because it states that the entity in charge of fostering the 
decentralization process was the National Council for Decentralization / CND, which 
no longer exists because it was merged with the PCM’s Decentralization 
Secretariat, turning it into an administrative entity." (Municipal organization) "The 
CND failed, that model failed and the PCM is insufficient. There is a pending 
agenda." (Legislative branch) 
21 
Lastly, respondents constantly refer to institutional parallelism:
"There are four types of organizations within the regional governments. One is 
from before the times of the CETARES, another is from the times of the CETARES, 
and then there are the regional directorates and regional management, so it’s 
very difficult to govern there. The regional health director reports to the minister 
of health. The social development manager does not have authority over the 
director of education, the Ministry of Education deposits money in the account of 
the Regional Directorate of Education so that it can prepare early childhood 
education profiles, the regional director of education doesn’t know what the 
money is for and the regional general manager doesn’t know that the regional 
government has received a deposit. That design doesn’t lead to good 
management." (Central government official). 
The annual report on decentralization, a tool that has been given periodically to key 
actors in the Peruvian government, sub-national government and civil society, has 
contributed to the debate and informed decision making. 
With regard to technical assistance on the issue of approval or reform of 
decentralization legislation, PRODES-II collaborated on the drafting of 11 bills, provided 
technical support on 8 initiatives on administrative modernization of regional and local 
governments, and supported 14 initiatives on decentralization by collaborating on the 
organization of debates and public forums. It should be noted that PRODES-II also 
identified some policies and regulations related to the decentralization process on which 
a high level of political will and consensus would be required: 
 Transfer of sectoral functions to regional governments. 
 Delineation of functions of different government levels. 
 Intergovernmental articulation and guidance of the decentralization process at 
the national level. 
 Concerted execution of the decentralization process. 
 Completion of the development of decentralized administrative management of 
sectoral functions transferred to Regional Governments. 
 Promotion of the definition of territorial boundaries. 
 Development of a classification for municipalities. 
 Promotion of the regionalization process 
 Promotion of national strategic planning as a national system. 
 Redesign of the fiscal decentralization model according to equity and 
partnership criteria. 
 Institutional reform and modernization of Regional Governments. 
 Development of a system for monitoring and evaluating results of the 
decentralization process. 
 Production and dissemination of relevant information by the Executive Branch. 
As in the preceding case, in this case the interviews indicate an on-target assessment 
regarding the needs of the decentralization process that should be taken into account in 
the new phase of the project. However, it should be noted that the respondents did not 
agree on the priorities of these pending tasks. On some points, there are conflicting 
positions, such as in the case of fiscal decentralization on which some analysts expressed 
their doubts: "With regard to fiscal decentralization, I think we should proceed calmly. 
22 
What we are discussing now is the weakness of regional governments for executing it, 
isn’t it? So I don’t see that there’s any hurry for regional governments do obtain their 
own resources." It can be concluded that there is a vacuum where a contribution to 
consensus building is needed among public policy developers, among the different actors 
involved in the project. 
A member of the Executive branch commented, "Yes, I believe that PRODES has made a 
very clear assessment; now it just needs someone to act on it." A similar position was 
stated by a Congressional functionary who said, "We have the information provided by 
PRODES.... the problem is that Congress needs to pay attention to it and take it into 
account." On this matter, a way of encouraging them to take action on it needs to be 
determined. 
In the qualitative performance evaluation of PRODES-II, questions were included to 
determine respondents’ perception of the project’s intervention in the development of 
the regulatory framework for decentralization. The findings considered most relevant 
are the following: 
 There is agreement among respondents at the national level when evaluating the 
materials produced by the program, especially the Decentralization Compass: "In 
general, I greatly appreciate Prodes’ work. I think they do good work; they’re 
very well informed." (Political analyst). However, it should be noted that the 
level of follow-up on the document is usually lower than the importance attached 
to it: 
"We’ve recently received a number of PRODES publications. Since it was just 
completed recently, we just found out about many of their books a short time 
ago. They sent us a complete collection of everything they’ve produced over the 
last 3 or 4 years. That’s how we found out about a set of documents that the 
mayors surely must have received directly. Apart from the mayors, we don’t 
know, but we do know that it contributed to capacity building and to making up￾to-date information available.” 
 Regarding the Decentralization Compass, the fact that its development had been 
coordinated with different actors in the Peruvian government, regional 
governments, local governments and other actors was particularly highlighted. 
Also assessed as positive was the recent transfer of the Congressional Commission 
on the Decentralization Compass, although the need to expand consultations 
made for its preparation to other regions of the country was mentioned. 
"The situation is that PRODES is over and we in Congress have to invent a new 
way to make the compass. We don’t know exactly how we’ll do it yet, but we 
have to make some inquiries outside Lima. The PRODES inquiries are 
concentrated on the departments where the project has worked, such as the 
high jungle and coca growing areas. We are going to have to do something more 
extensive, but we need to obtain financing and find different mechanisms to 
create the instrument in Congress." 
(Functionary of the Congressional Commission on Regionalization) 
23 
 In addition, the Congressional Commission on Decentralization highlighted the 
collaboration of PRODES-II, which provided technical assistance on the 
preparation of different laws relevant to the process that have been mentioned 
in performance indicators. In other words, both the documents contributed by 
PRODES-II and the process followed in their preparation were appreciated: 
The methodology is a consultation methodology and the compass was built based 
on consultations with various actors. Many inquiries were also made in the 
territories where the project intervened. We were part of the consultation made 
in Lima, but the result is an important tool to guide the work of the commission 
or members of Congress, a reference for the pending agenda, although it will 
not necessarily be dealt with." 
(Congressional functionary) 
 Respondents particularly highlighted the interaction workshops held with the 
participation of key actors, as spaces that offered a neutral platform where 
representatives of different areas of the Executive Branch (PCM, MEF, etc.) and 
the Legislative Branch (Congressional Commission on Decentralization), as well as 
sub-national governments and other relevant actors, could make contact and 
exchange views. Thus, one of the strengths of the project highlighted by 
respondents was the creation of informal coordination spaces. PRODES-II had to 
operate in a context marked by the absence of intergovernmental coordination 
mechanisms, overcoming the difficulty of reaching consensus and a lack of 
political foresight and leadership on the part of the Executive Branch and other 
key actors, due to which the on-target decision by PRODES-II to contribute to the 
ongoing process of articulation through the Interregional Coordination Boards 
(JCI) and to support the creation of municipal associations was highlighted. 
2.1.3 Aspects that limited or facilitated the program’s action on this point 
The aspects that limited the program’s action in this area were mostly inherent to the 
Peruvian political process, because as detailed in the chapter on the decentralization 
context, many of the regulatory reforms required depend on the political will of 
different actors, the development of a hegemonic vision and leadership of the process 
by the Executive Branch, and the ability to reach consensus and foster 
intergovernmental coordination processes that do not currently exist. 
In this regard, the accompaniment of PRODES-II on the development of the regulatory 
and policy framework had to be adapted to these limitations, collaborating only on 
aspects where there was consensus to move forward, in a context in which several 
respondents indicated that there was stagnation of the process: 
"The decentralization process is static; there is a lack of political leadership to 
move forward with the process, as well as a lack of technical tools. What one 
can see is that after the National Council for Decentralization was disbanded, 
the governing body, which is the Decentralization Secretariat, lost political 
weight. In addition, I think that from the point of view of the central 
government, the technical capacities to monitor the process are not adequate 
yet." 
(Former MEF official) 
24 
An important element to highlight is the low visibility of PRODES-II interventions in 
assisting with the development of the regulatory framework and information for policy 
development at the regional and local levels, where little is known about the 
Decentralization Compass and the Map of Regulations and Policies on Decentralization. 
This is due to the manner in which PRODES-II was designed; it should have involved 
regional actors more. 
Despite the lack of recognition of the work of PRODES-II in this area at the regional and 
local levels, the project’s on-target decision to contribute to the ongoing process of 
articulation through the Interregional Coordination Boards (JCI) and to support the 
creation of municipal associations was noted. Also mentioned was the foreign assistance 
provided to the National Assembly of Regional Governments (ANGR), although 
respondents felt that greater support was needed to coordinate and articulate proposals 
from regional governments and to give more prominence to the National Assembly of 
Regional Governments as a space for reaching consensus at the sub-national level and a 
link with the Executive Branch at the national level. 
2.2 Social Conflicts 
2.2.1 Objectives of the project 
In order to analyze the information obtained, it is important to note that PRODES II 
established the following as a general objective in conflict resolution: "To assist the 
Peruvian government in its efforts to design and implement a more effective strategy to 
prevent and handle conflicts through democratic channels and institutions. The 
assistance should increase public confidence in the government's commitment to 
resolving conflicts." 
The specific objectives of PRODES II regarding conflict resolution were: 
a. To assist the PCM on the development of a coherent, integrated vision and 
planning for conflict management, response and resolution, including 
contributions from key ministries and local officials in the areas where conflicts 
take place. 
b. To design an appropriate and effective communication strategy geared toward 
having representatives of the national and local governments, as well as civil 
society organizations, communicate actions related to the conflict resolution 
policy. 
c. To provide technical assistance to promote key changes in laws and policies 
thatfacilitates the implementation of a national conflict management strategy. 
d. To build institutional capacity in conflict management, prevention and resolution 
by training line ministries, regional governments and sub-national actors. Efforts 
should take place in the departments of San Martin, Amazonas and Madre de 
Dios, as well as other areas directly affected by conflicts. 
25 
Specifically at the national level, its work was related to the following: 
Legal and policy framework: 
a. Provide the PCM and the ministries’ most important conflict management offices 
with a more coherent and effective conflict resolution strategy that can be applied 
at the national and sub-national levels. 
b. Technical assistance on the preparation of a legal framework for conflict 
management in a joint effort with the PCM and key members of Congress 
committed to the issues related to conflicts (indigenous matters, the Amazon 
region, energy and mines, women, etc.), involving civil society in the debate. 
Institution building: 
a. High technical level meetings with the PCM’s conflict unit, the Prime Minister and 
key ministers, in order to develop a comprehensive outline for conflict prevention 
and administration. 
b. Design an information and communication strategy that includes mapping tools 
which will serve as a mechanism for analyzing and identifying conflicts. 
c. Preparation of management tools to raise the level of the Coordination Unit in the 
hierarchy to that of Conflict Management Office. (The ROF, MOF and CAP were 
approved.) 
At the sub-national level, the program’s intervention was to take place in the following 
manner: 
a. Prepare a training package on conflict management and resolution to be validated 
at the intervention sites. These packages should include tools and methodologies 
for conflict management and prevention. 
b. Provide training and technical assistance on capacity building for conflict 
management and prevention at the regional level that includes participation at the 
municipal level, community leaders, civil society and representatives of 
commercial sectors. 
c. Assistance on the implementation of a national policy that makes it possible to 
formalize the artisan and informal mining sector. 
2.2.2 The program’s contribution on the issue of social conflicts 
According to the Office of the Public Ombudsman, during the current administration’s 
term in office, as of July 2012 there were 247 on-going social conflicts, of which 150 
were related to environmental issues. According to information from the PCM, as of July 
of this year, the Social Conflict Management Office (OGCS) was working on more than 60 
social conflicts. 
26 
Regarding the context in which PRODES conducted its activities and will conduct its 
activities in the next phase, concern about social conflict is high among respondents; it 
is obvious that it is one of the issues in relation with which there is greater demand: 
"The state and companies have realized that they can no longer operate in the 
same way because when they build a road, install a transmission line, lay pipe or 
whatever, in some way it directly or indirectly affects a private or public 
interest. The level of awareness regarding participation and the capacity to 
express discontent is higher because of the media and education. Companies and 
public institutions are taking courses on how to assess conflicts and in my classes, 
because I teach at the Catholic University, I have many people from the Ministry 
of Transportation, the Ministry of Agriculture and the Ministry of Mines taking 
courses. They realize that they can no longer give the order that 'the highway 
goes there’ without having conducted a consultation process. By law - because 
that was another law that was passed - every government sector, of the state, 
must have a conflict management team.” 
(Specialist in conflict resolution) 
This confirms the importance of having included the issue and the need to incorporate it 
in the new phase of PRODES. From the point of view of a key actor, the work was 
performed in different areas: 
"…we took the capacity of this organization to support us into account in two 
lines of work: the first line is the design of the office of social conflict 
management. With the contribution of PRODES and the consultant they hired, 
what was to be a new office with a new approach to social conflict management 
was designed. Work on this output was performed with the PCM’s office of 
public management. The result was a proposal for the reorganization of this 
office. The second line of work was training on social conflicts for regional 
government officials and representatives of institutions in the sectors involved 
in managing them." 
As usually occurs with project interventions, the training component received 
recognition: "PRODES has been holding training workshops. Training materials were 
developed for three modules and 200 to 300 people were trained on conflict 
resolution. Participation was excellent, but more work is still needed.” 
(Specialist in conflict resolution) 
 From the information gathered through interviews, the following can be established: 
 PRODES’ contribution to having seven regions with greater insight on the subject 
of conflicts is acknowledged. This step requires accompaniment to consolidate its 
achievements: "There are seven or eight governments that now have a broader 
perspective regarding conflicts and the possibility of working on them, but how 
to go about it, the matter of establishing conditions, relations, all these sensitive 
issues are not being dealt with yet.” (Specialist in conflict solution) 
 The importance of having provided training is acknowledged. However, it was 
also mentioned that the effects of this training are reduced when the way in 
which each level of government involved should intervene is unclear. "I believe 
27 
that Regional Governments need on-going training. A workshop is not enough. 
They need more support. "(Specialist in conflict resolution) 
 Training is limited by staff turnover, as mentioned before. There are cases, such 
as Ayacucho, where officials have a low level of knowledge of the training 
programs held. 
 PRODES II not only provided training; among its achievements are the approval of 
the Guidelines and Strategies for the Management of Social Conflicts (2011) by 
means of a Ministerial Resolution. However, the effect of this progress is not yet 
clearly perceived: "On the issue of conflicts, legislation was passed, specific 
legislation on conflicts, in which we proposed functions and conflict 
competencies for the national government. This conflict management office 
that I mentioned, and also for the regional governments with more specific 
capacities, competencies and functions for regional governments. The 
problem was that the administration in office at the time, the APRA 
administration, cut several of those functions. In the end, it seems that in 
the last period before the PRODES decentralization program ended, work 
was done on legislation on the competencies of regional governments in 
relation with the issue of conflicts, so that draft legislation was prepared 
and presented to the office. I don’t know what its status is now. (specialist on 
conflicts), "What I know about is the design of the Directorate of Social 
Conflicts that served to place it on the agenda of the PCM and the office in 
charge of PCM programs, and it ended up in a proposal that’s being 
discussed.” (former central government official) 
 The Geo-referential Information System and the website of the Conflict Office 
have been developed and monthly newsletters have been published, although 
they were not mentioned by the actors consulted, which may have to do with the 
fact that pertinent decisions have not been taken by the central government. 
 The National Plan for Formalization of Artisanal and Informal Mining, which is 
currently in force as the framework for the process of formalizing illegal mining, 
is in line with a position indicated repeatedly by those consulted. Dealing with 
the issue of social conflicts not only requires working on consultation processes, 
but also focusing on the root problems, which requires an analysis of the 
situations that lead to conflicts. This type of intervention should be extended to 
other issues that give rise to conflicts in each region, such as water, natural 
resource management in general, etc. 
2.2.3 Aspects that limited or facilitated the program’s action on this point 
Intervention on the issue of conflict involves working on an extremely broad and 
complex matter, where it is necessary to clarify the agenda and work at the nation and 
regional levels to find an appropriate way to analyze social conflicts, design intervention 
processes, and manage early identification, monitoring and evaluation of social 
conflicts. Similarly, a system for monitoring progress on these matters is needed. 
The first matter to consider is that there is no clear assessment of the reasons why 
conflicts arise or why they spread. According to some, "The problem of social conflict 
28 
is not due solely to a lack of skills or a strategy for preventing and managing 
conflicts; it also has a significant political component.”(Congressman) Others 
attribute the problem to a lack of political parties and organizations: "I see that the 
actors in confrontations have carte blanche to spark crises. I think one of the most 
serious problems existing in the country is the lack of democratic leadership, the 
lack of organizations." (Former head of the Conflict Office) 
It was mentioned that the lack of political parties leads to local and regional authorities 
being "abducted" by unsatisfied social demands. They explain that in the absence of a 
political party that supports the authorities, they can only opt to take a stand in favor of 
these demands or lose their legitimacy. As a result, their role as mediators or conflict 
resolvers is diminished. Along the same line, local authorities are concerned about 
conflicts motivated by the pursuit of power: 
"We have many social conflicts. We have conflicts due to interest in gaining 
power, the desire to become part of the local government. There are those who 
do not run for mayor or participate in the elections, but want to be mayors. 
Sometimes one does not define those accompanying one, and it turns out that 
the deputy mayor wants to be mayor and organizes the population, generating 
instability through social conflicts." 
 (District mayor) 
In short, the issue has a political component that PRODES is not responsible for dealing 
with, but that it must take into account in the development of institutional redesign 
strategies directly linked to the decentralization process. 
When evaluating the position of the central government on the issue of conflict, 
respondents were critical: "It is not well prepared; its awareness is very general. It’s 
sensitive when its projects are stopped, when it costs it votes and popularity, but 
many of these conflicts need a real commitment to change, to taking another 
approach. You can have these conflict offices, but they must be implemented. 
There is still a lot to be done, but apparently the government does not yet see it as 
a priority. I think progress is being made on some matters; for example, there is 
talk of a dialogue office and it looks like they are going to reorganize the conflict 
management office, but I still don’t see a change in the approach promoted by 
these different sectors. I don’t see it yet.” (Specialist in conflicts) “We have a state 
that is not capable of preventing or foreseeing conflicts. I think that in general, we 
have a state that does not have this issue among its priorities.” (Specialist in 
decentralization) From this point of view, despite the fact that the central government 
is requesting that work be done on the issue of conflicts, it does not appear to have a 
clear idea of what is needed, which may cause complications in the logic of a demand￾based intervention. 
The necessary link between the central government and regional governments is 
extremely fragile. For this reason, one person commented, "What I do know is that the 
when the National Government has a conflict with a Regional Government, it 
freezes its accounts. They fight each other and articulation goes down the drain. As 
this is a very political issue, any type of articulation or coordination depends 
heavily on political will, the context, the current situation and whether or not they 
29 
have a shared vision" (specialist in decentralization). In this context, training or 
capacity building does not appear as a central theme. Efficient coordination mechanisms 
must be developed first. 
In addition, it was repeatedly stated that the role of regional governments is not yet 
defined. Currently, they are unable to make decisions about many of the matters that 
generate conflicts (the SUTEP strike was given as an example), a fact that contrasts with 
their impression that mechanisms to control the performance of presidents of regional 
governments are lacking. From this perspective, we are far from being able to take a 
position such as the one stated by a specialist: 
"To the extent that decentralization takes place, primarily the decentralization 
of conflict management, officials at the local, regional or provincial level will 
have more confidence to be able to intervene in these conflicts, without saying, 
‘It’s not my job’. They will clearly understand that it is their responsibility to 
prevent conflicts. " 
We can conclude that the intervention is marked by the importance of the issue, but 
consensus must still be reached on the role of each actor and how actors should interact 
at different levels of government. 
2.3 The contribution from a gender perspective 
With regard to the implementation of the gender approach, it can be found in the 
application of the activities executed by PRODES-II. Training on was held on female 
candidates in decentralized public management, in addition to training programs for 
male and female council and assembly members. A review of the attendance lists for the 
training programs shows that the objective of having more than 30% female participation 
was achieved. 
These workshops were intended for participants to take gender into account, as well as 
to build capacities for serving in the positions to which they were elected or for which 
they were running. According to the data collected from the program itself, those 
participating in the training program prepared 53 projects during the period between 
2009 and 2011, 35 of which included a gender approach. The above mentioned figure 
covers some constituencies that were not involved in this study; in Ucayali, Ayacucho 
and San Martin, 11 such projects were recorded.However, this was not the only 
achievement of the project, as indicated by respondents. 
From the information provided, it can be noted that the content of the training 
programs reinforced the gender issue and how it has helped the women who participated 
in them to develop a clear perspective on the issue and incorporate it in their projects. 
"I’ve seen that all of the training programs organized by PRODES taught that 
women have the same capacities as men, both to become involved in politics and 
to serve in a public function in municipalities or any other public institution. 
They teach participants that women can hold any position they want and have 
the right to do so, provided they are properly trained. I personally feel that 
PRODES has given me primarily academic training, and this training has made me 
more sure of myself and more aware that I can do anything I set out to achieve, 
30 
both politically and in the pursuit of development and prosperity for my 
community.” 
 (Municipal functionary in Puquio) 
"What we have seen in PRODES’ activities is that equality between men and 
women in the civil service has not been achieved. It does not exist, so we must 
work toward equality. Based on what we have learned in PRODES, we have 
developed a small intervention program in our own regional government, 
consisting of training both men and women to improve on the execution of the 
specific functions they perform and civil service itself. With this, we want to 
ensure that the women currently working in our regional government do not feel 
displaced in favor of male staff. We want women to become aware of the 
importance of their role in both public administration and politics in general." 
 (Council Member, San Martin) 
"Well, in its training sessions and in every program supported by PRODES, a 
bigger role for women has always been sought. PRODES has accompanied us in 
the implementation of participatory budgeting. Through its participation, the 
participation of committees of female entrepreneurs, female board members of 
wawawasis (day care centers) and most women's organizations based in Ucayali 
has been encouraged. PRODES believes that the participation of women in 
politics and in the development of their community should be promoted, which 
is why it is committed to building their capacities to achieve both objectives." 
(Municipality of Padre Abad)
A finding worth mentioning is that awareness of the gender perspective is raised not only 
by organizing activities exclusively for women; good results can be obtained in training 
sessions held for male and female officials at the same time. 
It should be noted that the training program participants prepared 38 draft proposals for 
Regional Ordinances. Of the total number of Ordinances prepared, 15 included the 
gender approach, and 02 of the latter had been approved as of June 2012. If we focus 
only to the area of the influence of the current research; the regional government of 
Ayacucho approved the Regional Ordinance N°021-2011 GRA/CR, which declares as a 
public policy the improvement of public maternal health, being pending of approval: 
“The internship program for elementary school students from Ayacucho rural areas” and 
other refer to address the issue of school dropouts in the rural areas of elementary 
school students. 
31 
Capacity Building
Two approaches are presented in this part of the report. The first one is the quantitative 
evaluation of the project’s performance and the second one is the analysis derived from 
the interviews conducted. A third point refers to the main findings derived from the 
integration of both perspectives. 
3.1 Objectives and indicators of the capacity-building program 
As part of the overall objective of the PRODES program, capacity building for more 
effective governance was planned for selected Regional and Local Governments (sub￾national strengthening). To this end, training was provided on diverse matters to support 
the implementation of national policy at the regional and local level, as well as to 
develop their capacities further. 
The program focused on providing training and technical assistance to four regional 
governments and about 300 municipalities. According to the Catalogue of Resources and 
Materials for the USAID / Peru PRODES Project2, training was provided for public 
authorities and officials of regional and local governments for the purpose of 
contributing to improved performance of their functions, and civil society capacities 
were strengthened for effective participation in participatory public management. The 
capacity building programs included a training program for male and female regional 
council members, a training program for male and female municipal assembly members, 
a certificate program on the formulation of public investment projects with a human 
development and child welfare approach, a training program for investment promoters, 
a training program for community leaders, and workshops on results-based budgeting 
and management, contracting, procurement and strategic planning, among other public 
management topics. Technical assistance was provided on specialized topics such as the 
implementation of the Unified Administrative Procedures Text (TUPA) model, municipal 
tax collection and participatory budgeting, among others, based strictly on each 
government entity’s needs. 
3.2 Quantitative research 
As the PRODES Project did not include the possibility of rigorous assessment of the 
project’s impact, USAID included both the qualitative and quantitative approaches in the 
design of the final evaluation. In this regard, the quantitative approach will be limited 
to some factors of the program’s intervention on which the data has made rigorous 
analysis possible. Therefore, it was possible to review secondary information from 
reliable and GOP official sources. With the consent of USAID, the Evaluation Team 
proposed focusing the quantitative evaluation on two indicators (Y variables) to analyze 
the PRODES Program’s performance in the regional governments of Ayacucho, San 
Martin, Ucayali and their respective provincial municipalities. These indicators were the 
following: 
a. Effectiveness in the execution of the investment budget 
b. Number of public investment projects approved 
Chapter 3
32 
First, the impact of the presence of the program as a whole is evaluated, and then only 
the effect of the number of training courses and the number of beneficiaries is 
considered. Given the heterogeneity of the technical assistance provided and the 
heterogeneity of this type of intervention, they will not be considered in the second 
phase. 
These indicators were selected due to the fact that they coherently show whether the 
expected outcome of the program at the sub-national level, which was strengthening of 
the capacities of sub-national governments to govern3
, has been achieved. In addition, 
they are indicators of the outcome established by PRODES in the Performance 
Management Plan, in addition to being those best adapted to the quantitative 
methodology to be used. 
The number of public investment projects approved by the SNIP (National Public 
Investment System) is an indicator of effectiveness in their formulation and approval. 
The series that show the projects approved for government entities reveal dissimilar 
results over the period of analysis. In regional governments, there appears to be a drop 
in the number of projects, with the exception of Ucayali, while in the case of provincial 
municipalities, the series appears to have a positive slope. In general, no results can be 
inferred from this descriptive analysis when comparing the regions intervened with a 
similar region (control group). 
 
3
Terms of reference, outcome 2 (p. 7).
33 
However, the reduction in the number of projects approved does not imply that the 
amount invested in each one was lower, due to which the indicator is only referential. 
For example, in the case of the Regional Government of Ayacucho, the number of 
feasible projects from 01/01/2009 to 12/31/2009 was 157, with an investment amount 
of 274.9 million nuevos soles, whereas from 01/01/2011 to 12/31/2011, the number of 
feasible projects was 62 and the amount invested was 338.8 million. The projects in 
which the largest amount was invested were agricultural sector projects in both years. 
Source: Governance Barometer
Preparation: In‐house
Graph No. 1: Approved Public Investment Projects
2008-2011 
34 
In the case of the Regional Government of Ucayali, the number of feasible projects from 
01/01/2009 to 31/12/2009 was 71, with an investment amount of 188.5 million nuevos 
soles, whereas from 01/01/2011 to 31/12/2011, the number of feasible projects was 67 
and the amount invested was 213.3 million. The projects in which the largest amount 
was invested were in education, culture, and sanitation4
. It should be noted that the 
fact that a project is approved does not necessarily mean it is executed. In addition, 
there are criteria for establishing a project’s priority in relation with others. In this 
regard, there is also a trend among regional governments to approve large-scale 
strategic projects. 
With regard to the effectiveness of budget execution, the indicator measures the 
investment budget executed (accrued) over the total modified institutional budget (PIM) 
for investments for each year of analysis. In general, trends show a drop for all 
government entities in 2011, whether intervened or not. It should be noted that 2011 
was the newly elected regional and local authorities’ first year in office, and 
presidential elections took place in July. As it was an election year, according to law5
 , 
non-financial government spending executed during the first seven months of the year 
was not to exceed 60% of the non-financial expenses budgeted. However, it appears that 
that in the government entities in which the program intervened, the decrease in the 
budget executed is less than in the control group, especially for provincial 
municipalities. 
 
4 Source: SNIP Project Bank. At constant 2009 prices (INEI). For some years, the number of feasible projects
in the Project Bank does not tie to the number of feasible projects according to the Governance Barometer,
due to which the number from the Project Bank is being used only in this case.   5 Law No. 27245, Art.4.
Graph No. 2: Level of budget execution
2008-2011
35 
3.3 Training and beneficiaries 
The total number of training sessions and beneficiaries is shown in Graphic No. 3 for every topic 
selected for capacity building through the program. For the Ayacucho region, the largest number 
of training sessions is in the area of institutional reform and modernization with 29 training 
sessions. However, there are greater numbers of beneficiaries in the areas of management and 
results-based budgeting and results-based participatory budgeting with 570 and 562 individuals 
respectively. For the San Martin region, the largest number of training sessions is in results-based 
participatory budgeting with 43 training sessions, and strategic planning with 28 training sessions. 
The largest number of beneficiaries is in the area of results-based participatory budgeting with 
3109 beneficiaries, followed by strategic planning with 1123 beneficiaries. Lastly, in the Ucayali 
region, results-based participatory budgeting leads with 33 training sessions and 1242 
beneficiaries, followed by strategic planning with 573 beneficiaries and access to public 
information with 18 training sessions. 
In general, the greatest demand corresponding to regional governments is concentrated in the 
area of results-based participatory budgeting, with a total of 2927 beneficiaries trained and 105 
training sessions. 
Fuente: Barómetro de Gobernabilidad 
Elaboración: Propia 
36 
Graph No. 3: Number of trainings and number of beneficiaries 2009 - 2012
It should be noted that the number of training sessions and beneficiaries largely 
corresponds to the demand from each government entity and the target public. For 
example, training sessions and technical assistance on topics related to the SNIP are 
geared toward very specific personnel, due to which the number of beneficiaries does 
not determine the full impact of the proposed indicators.
Data estimation methodology, treatment description and management 
The methodology used is the panel modeling approach, given the complexity of the 
observation units and the availability of data for each government entity, following 
Angrist and Krueger (1998). According to Benavente (2004), a set of panel or longitudinal 
37 
data is that which follows individuals (firms, governments or countries) over time and, 
consequently, provides multiple observations for each individual. Thus, panel data 
combine time series and cross-sectional data. 
One of the advantages of this model is that it allows for correcting important 
econometric problems associated with omitted variables or unobservable effects, which 
lead to biased and inconsistent results. That is to say, the panel makes it possible to 
control for unobservable factors that vary across individuals but remain constant over 
time or vary over time but remain constant across individuals. 
A specification of the panel data model is the fixed effects model, which takes the 
following form: 
Yiti TRitXitdtDteit
 (a)
Where subindex i represents each entity and t each year analyzed, Yit corresponds to 
indicators to be assessed (investment budget execution and number of PIP approved), 
Xit variables correspond to observable characteristics of the units analyzed, TRit 
corresponds to the variable and its coefficient, is the parameter to be estimated, 
which represents the impact of treatment, i denotes unobserved cross-sectional 
heterogeneity, Dt are dichotomous variables for each year analyzed that serve to 
introduce temporary, unobserved heterogeneity, and eit is a random variable that 
represents error in estimation. 
Thus, in the specification (a) we are assuming that the unobservable cross-sectional and 
time-varying factors are absorbed by the constants i and y dt respectively, and not by 
the error (i.e., we assume that they are not random, but rather fixed). This model assumes 
that individual effects are deterministic parameters which must be found. This is the case 
where N represents a "population" of government entities, and the inferences relate only to 
the N available. Thus, the inference is conditional on the individual N and not other sets. 
Firstly, the treatment has been assumed to be the effect of the program as a whole on the 
regions intervened without directly differentiating between training and technical 
assistance. Therefore, a dichotomous variable with a 0 value is specified if the government 
entity did not receive the program and 1 if it did. 
The second manner of describing the treatment is to assign the total number of training 
sessions or total number of beneficiaries to this variable: 
‐ Number of training sessions (t): Number of training events held during the 
year.
‐ Number of beneficiaries (t): Total number of people participating in training 
events per year. 
TRi
38 
These variables have been calculated at the departmental level (including provincial 
municipalities and regional governments), as the information available has not allowed 
adding provincial-level municipalities and regional governments independently. 
Therefore, aggregation is still perfectible. 
In addition, five training topics were selected from the total list of topics6
, which can 
influence the indicators to be evaluated in a more direct way, thereby obtaining two 
more variables: 
‐ Number of training sessions (5): number of training events per year, 
corresponding only to the five topics selected 
‐ Number of beneficiaries (5): Total number of people who participated in 
training events, corresponding only to the five topics selected
As mentioned previously, only the training sessions will be considered in this phase of 
the evaluation. As technical assistance services are of a very specific nature, they are 
much more heterogeneous in terms of their type, length and frequency, not only from 
one intervened region to another, but also from one government entity to another in the 
same department. This means that the proposed methodology does not show credible 
results. That said, the total number of technical assistance services is shown in Chart 
No.A-2 from the appendix. 
However, despite being more homogeneous than technical assistance services, training 
sessions have a set of characteristics that should be mentioned. It should be noted that, 
due to the project’s demand-based approach, the frequency of training sessions also 
varies; in addition, there is a certain relation with the number of provinces, the size of 
the population in each region intervened and the type of training each one requires. 
Adding training courses lasting different lengths of time and on different topics is not a 
futile effort. In addition, the database prevents determining whether the same 
individual participated in more than one training program. Therefore, the aggregations 
are only indicative and the results are can only be interpreted in the sense of the sign, 
of the statistical correlation between the intervention variable and the outcome 
variable, and not the magnitude of the coefficient. (A description is presented in Chart 
No.A-1 in the appendices). 
The control group for each government entity comes from a baseline study 
commissioned by the program in 2009.7
 For the control group of regional governments, 
the variables used were total population, number of provinces, percentage of rural 
population, percentage of urban population, per capita budget, and natural region 
(coast, highlands or jungle). For provincial municipalities, the variables used were total 
population, percentage of rural population and per capita budget. 
This group was useful for a preliminary exploration of quantitative analysis in which the 
dichotomous variable was used as the treatment variable (whether or not the entity 
 
6 These five topics are: Key aspects of public management; training on results‐based management and budgeting;
contracting, procurement and public works; formulation and evaluation of Public Investment Projects; and
administrative simplification. 7
Baseline of the ProDescentralization program, governance barometer and composition of the control group.Final
report. Patricia Zárate, September  2009. Detailed in Table No.A‐1 in the appendices.
39 
received the program). However, in the case of regions, the amount of data is quite 
small; it covers only 24 observations (three beneficiary regions with their respective 
controls, for 4 years). In addition, a test of means was made (Table No. A-2 in the 
appendices), which shows that in most cases, the null hypothesis of no difference 
between the means of the treatment group and the control group is not rejected. 
For the second treatment description, where the dichotomous variable was replaced 
with the number of training sessions and number of beneficiaries, the database was 
modified to make it consistent with the methodology. The treatment variables were 
calculated at the departmental level, without making a distinction between regional 
governments and provincial municipalities, since the data did not allow differentiation. 
The X variables (population, percentage of urban population, percentage of households 
with water, percentage of households with electricity, percentage of illiterate 
population) are also at the departmental level. The number of public investment 
projects approved is the average of all the provinces in the department. Similarly, the 
budget execution percentage corresponds to the execution of all provincial 
municipalities in the department. This may have implications for the accuracy of the 
results because, on the one hand, there is the treatment variable at the aggregate level 
of both regional governments and provincial municipalities (total of all training sessions 
and beneficiaries in every geographical space) and, on the other, the Y variables 
correspond only to provincial municipalities. 
Lastly, a test was performed on this second treatment description with a more 
homogeneous sample, eliminating from the treatment group the departments with the 
greatest differences in certain observable characteristics from the previous sample. For 
this purpose, the Propensity Score technique, which calculates the probability of 
belonging to the treatment group, was used, building a Common Support and 
automatically eliminating departments outside the distribution. The variables used in 
the Propensity Score were population, percentage of urban population, and percentage 
of households with at least one unmet basic need. The Common Support made 
comparing the groups possible, but at a cost of reducing the sample; only 10 
departments remained. For this reason, the sample was reduced from 96 to 40 
observations (10 departments for 4 years).8
The data used in this study is from various sources for the 2008-2011 analysis periods. 
The number of public investment projects approved and the execution of the investment 
budget (Y variables) were obtained from the Governance Barometer Management 
Indicators Database. The list of training sessions and the number of beneficiaries was 
also provided by PRODES (treatment variables). 
The National Household Survey (ENAHO) for the 2008-2011 period provides the 
percentage of households with at least one unmet basic need, the percentage of 
households connected to the public water network, the percentage of households with 
electricity, the percentage of urban population and the number of years of formal 
education for each geographical area corresponding to the regional governments and 
provincial municipalities. Data on mining royalties, transfers received by regional and 
local governments from the central government from mining tax payments from 2008 to 
2011, were obtained from the query module on transfers to regional and local 
 
8 The balancing of Common Support is shown in Appendix 3.
40 
governments on the website of the Ministry of Economy and Finance. Lastly, data was 
obtained from the Population and Housing Census of 2007, from which the total 
population in 2007 and the percentage of urban population in 2007 were used. 
Limitations of quantitative analysis 
The main limitation of quantitative analysis is that this project was not designed for an 
impact assessment, due to which there is no rigorous baseline or monitoring of 
indicators suited to this methodology. 
In addition, there are limitations as to the specifications of the model. It should be 
noted that what is being modeled are government entities and not individuals; there are 
non-observable variables that are characteristic of each region. Also, there are 
limitations to evaluation of the demand-based approach. While it is believed that it is 
more efficient to provide training to the regions according to their needs, this makes the 
evaluation very endogenous. It is the same government entity requesting the 
intervention that ultimately is not consistent across the sample. This heterogeneity in 
the treatment makes it more difficult to evaluate by causing potential statistical 
problems such as endogeneity. In it, explanatory variables could be correlated with the 
error term; there is a bias when determining the causal relationship between Y variables 
and X variables. Also, it is necessary to detail the reasons why the regions selected were 
intervened, in order to make it possible to control for these effects and extrapolate the 
results to other similar regions. 
For example, it was not possible to observe the quality of the training, which certainly 
has an impact on the outcomes to be measured (however, according to the information 
gathered in interviews, the quality of these training sessions was very good). 
When aggregating by number of training sessions, their number in each geographic 
region was taken into account, measuring the impact of all of the training sessions 
separated in 5 topics. It should be noted that the presence of more training programs 
held by the state or other foreign assistance organizations has not been modeled; nor 
has the effect the elections may have had on the outcome variables. Similarly, it was 
not possible to determine whether the same individuals participated in more than one 
training program or if they stopped working for public entities as a result of their 
improved capacities. 
Additionally, there is a lack of information on the characteristics of each regional 
government and provincial municipality. With such information, it would be possible to 
build a more homogeneous control group and therefore control for more variables, such 
as percentage of professional employees, total number of employees, equipment, 9
 
number of computers, etc. 
It should be noted that, while the increase in approved investment projects has an 
expected correlation with the capacities of a limited set of officials, the result regarding 
the execution of the investment budget involves the capacities of a large group of 
officials (those in charge of spending, procurement, logistics, etc.) and other 
institutional variables that the model does not capture. 
 
9
Information contained in the RENAMU survey, available only up to 2009. 
41 
The quantitative study refers only to variables susceptible to an impact evaluation, 
which do not reflect the full scope of the intervention of PRODES. In turn, the treatment 
variables used do not reflect every action taken by the program. Technical assistance is 
not being taken into account and it is known that the number of training sessions and 
the number of beneficiaries correspond to the demands of each government entity. 
Despite these limitations, the econometric estimates give indications of the program's 
impact on the indicators evaluated. 
Estimation results 
Chart No. 1 shows a summary of the estimates, mainly the estimated parameter of the 
program’s impact on the proposed indicators. The number in parentheses represents 
statistic t, which indicates whether or not the parameter is statistically significant. Over 
1.28, it indicates that the parameter is statistically significant at 20%, 1.64 or over at 
10%, 1.96 or over at 5%, and 2.57 or over at 1%. R2 is a coefficient between 0 and 1 that 
measures whether both proposed X variables explain the model as a whole. Lastly, N is 
the number of observations in the regression. 
Approved Investment projects 
In a preliminary approach, taking the group established in the baseline study as the 
control group and the intervened regions as the treatment, the results show a positive 
impact on provincial municipalities for approved public investment projects at a 
statistical significance level within 20%. In the case of regional governments, the impact 
is statistically significant. It should be noted that the number of observations is very 
small in this case. (See also Table No. A-4 in the appendices) 
In a second approach, when using the number of training sessions as the treatment 
variable10 and considering the total number of departments, the impact on the number 
of approved projects indicator is positive and has a significance level of about 15% (see 
Table No.A-5 in the appendices). These results are consistent with those obtained when 
replacing the treatment variable with the number of training sessions on only 5 topics 
(those most closely related to the indicator evaluated), where a positive impact is also 
observed, and there is a higher significance level of 10% (see Table No.A-7 in the 
appendices). The results are similar for the number of beneficiaries. 
Lastly when restricting the previous analysis to observations within Common Support,11 a 
positive impact is observed at a significance level of 1% (See also Table No.A-9 in the 
appendices). The total number of observations in this case is 40. 
The impact and statistical significance level improve when the treatment variable is 
changed to training sessions covering only the five topics most closely related to the 
indicator to be evaluated, as occurs when the sample is restricted to a more 
homogenous one. 
 
10All of the variables at the regional level have been aggregated in this case. 11As explained in the Quantitative evaluation of the project’s contribution section, the sample was restricted to a more
homogenous one, in order to make the treatment group more comparable to the rest of the departments.
42 
Budget execution 
In a preliminary approach, taking the group established in the baseline study as the 
control group and a dichotomous variable related to the presence of the program as the 
treatment variable, the results show an impact close to zero, but statistically 
insignificant for provincial municipalities, and for the regional governments, a similar 
impact at a significance level of 10%. However, the number of observations is very 
small for the regions. (See also Table A-4 in the appendices.) 
In the second approach, when the number of training sessions was used as the treatment 
variable, the impact of the program on the budget execution indicator was almost zero 
and statistically insignificant. (See also Table A-5 in the appendices.) These results 
remain the same when the number of training sessions on only 5 topics is used and when 
the sample is restricted to those of common support (see also Tables No 7 and 9 in the 
appendices). These results are similar for the number of beneficiaries. 
Chart No. 1: Summary of the Main Regressions 
Sample: Treatment and Control Group (Baseline) 
 % Budget Executed 
(Regions) 
% Budget 
Executed 
(Provinces) 
No. of PIP 
approved 
(Regions) 
No. of PIP 
approved 
(provinces)
Treatment 
(dichotomous) 0.227 0.014 -23,029 7,105 
T (1.65) (0.33) (1.21) (1.43) 
R2 0.32 0.03 0.78 0.04 
N 24 200 24 200 
Sample: All Departments in the Country 
 % Budget Executed 
(Provincial Average for the Dept.) 
No. of PIP approved 
(Provincial Average for the 
Dept.) 
# Training sessions -0.001 0.344 
T (0.76) (1.85) 
R2 0.14 0.08 
N 96 96 
# Beneficiaries 0 0.008 
T (0.85) (1.74) 
R2 0.14 0.07 
N 96 96 
# Training sessions (5 
topics) -0.004 1.373 
t (0.77) (2.06)* 
R2 0.14 13 
N 96 96 
# Beneficiaries (5 
topics) 0 0.024 
t (0.71) (1.22) 
R2 0.14 09 
N 96 96 
 
43 
Sample: Departments within Common Support 
 % Budget Executed 
(Provincial Average for the Dept.) 
No. of PIP approved 
(Provincial Average for the 
Dept.) 
# Training sessions -0.001 0.543 
t (1.22) (3.11)** 
R2 0.6 0.57 
N 40 40 
# Beneficiaries 0 0.014 
t (1.66) (3.27)** 
R2 0.62 0.58 
N 40 40 
# Training (5 topics) -0.008 1.804 
t (2.28)* (2.93)** 
R2 0.66 0.55 
N 40 40 
# Beneficiaries (5 
topics) 0 0.038 
T (3.42)** (1.92) 
R2 0.72 0.46 
N 40 40 
* p<0.05; ** p<0.01 
The different estimates made show signs of a positive effect of the program on the 
number of public investment projects approved at a statistical significance level 
between 1% and 20%, which is acceptable considering the number of data. Parameter  
is positive and statistically different from zero. The statistical significance level 
improves as the treatment group becomes more comparable with the control group, as 
observed for estimates using a common support. 
Regarding the impact on the investment budget performance indicator, the methodology 
does not yield conclusive results. One possible explanation may be that the effect on the 
budget is more difficult to measure, as it not only depends on the capacity of 
governments, but also on other factors such as the complexity of the National Public 
Investment System (SNIP), the mechanisms, difficulties and delays in contracting and 
procurement processes, eventual delays in government coordination in making resources 
available to the executing units, and other logistical problems in general. In addition, 
the variable does not capture other interventions through which PRODES and other 
USAID, MEF and other donors’ projects support improvement in public management in 
these same environments. It should also be noted that the majority of regressions only 
include the number of sessions, and not the number of technical assistance services, 
which were an important part of the intervention. 
3.3 Qualitative evaluation of capacity building 
 
Training is the component of PRODES best recognized by respondents, not only in terms 
of recollection and identification, but also due to how it is valued. One of the aspects 
that stands out is that this attitude is found among those who have been allies of 
PRODES in conducting training sessions (such as UNICEF or MEF), as well as opinion 
leaders, authorities and beneficiaries (regional council members, for example). 
44 
The menu of options developed by the entity evaluated is quite extensive in terms of 
target publics and topics. The training of council and assembly members, which had the 
added advantage of being carried out in coordination with government and civil society 
entities, as well as other donors, was mentioned. Also mentioned was training on project 
development and the Unified Administrative Procedures Text (TUPA). This forms the 
profile of an entity that ranges from fairly generic to specialized topics. What is the 
specificity of PRODES in such a wide-ranging offering? Respondents clearly recognize 
that the it has developed a very efficient working methodology that has even been 
adopted for other training processes (this is mentioned especially by PRODES allies; i.e. 
entities that participated in the design of the training programs). Allies of the program 
see it the way a person interviewed by the MEF does. This person commented that 
PRODES has "a well-known training model that has been disseminated by international 
foreign assistance. Thanks to it, we had successful programs." 
With regard to entities that worked with PRODES, they repeatedly mention how the 
alliance enabled them to achieve their objectives. This is the point of view of the Office 
of the Public Ombudsman ("Our main experience with the Prodescentralization
project was the training program for regional council members”) and non￾governmental organizations such as Prisma (“PRODES’ work with the governance and 
transparency project we executed enabled us to reach 19 regions with the course 
for council and assembly members”). The same is true for UNICEF, whose objective of 
offering regions a different type of training was easily aligned with PRODES’ objectives. 
According to comments received, PRODES appears to have contributed to the experience 
of others, as in the case of UNICEF; to them, PRODES’ contribution to the human 
development component is clear. 
 "We were just looking at this matter of capacity building among officials at the 
regional and local levels. Then we said, ‘Well, we have to do something to help 
them formulate public investment projects,’ because that was one of the major 
weaknesses. They did not have all the skills required to formulate public 
investment projects, particularly on social issues, with an approach centered on 
individuals. What we did do was incorporate the human development approach 
that we call person-centered. So we reviewed everything they had done, because 
we said we were not going to reinvent the wheel; there is already something 
that has been done successfully. They have a methodology geared primarily 
toward adults that is not the methodology where you sit in front of the class like 
a traditional teacher, but rather it is learning by doing.” 
In cases such as the foregoing, it is clear that there was a synergy that helped build 
public officials’ capacities and established a different way of conducting training 
workshops. 
When analyzing characteristics of the training activities, the quality of the professionals 
who led the training sessions and the quality of the materials are basically what 
respondents mentioned. With regard to the latter, there is fairly broad consensus. Some 
even consider it as one of the program’s most important contributions, as shown by the 
answers given by the Prisma representative who said, "I acknowledge PRODE’s ability to 
45 
produce materials; they have produced many materials that have been important in 
local capacity building.” 
From the beneficiaries’ point of view, those who received training acknowledge the relevance of 
the course, the appropriateness of the topic and especially the results obtained from it. In the 
case of training for council and assembly members, there are areas where differences with 
regard to previous changes of administration are mentioned, as in the case of Ucayali, where 
representatives of the regional government said, "Well, I think that members of the regional 
council have become aware of what decision making means. As a result, the work of the 
Regional Council has improved. Unlike last year, some assembly members have formulated 
draft ordinances, and that contributes to governance." Similarly, in Ayacucho it was noted 
that "What has made an impact is the training program for male and female assembly 
members ... because the invitation was open and many female assembly members who had 
never had access to any information on the decentralization process, their legislative 
functions, representation or oversight, participated. Before that, they were doing the best 
they could." 
While these are not results that can be verified quantitatively, it should be understood that, as 
indicated by the respondents themselves, capacity building requires time to take full effect. This 
would indicate that this type of training is effective, but its impact depends on the level of 
training elected officials had beforehand (the lower it is, the greater the potential for a 
perceptible difference). 
With regard to training on the TUPA, significant results were also reported in the provincial 
municipality of Huamanga: 
 "For example, we established our TUPA in a traditional manner, with a cost. A 
birth certificate would cost ten soles for no reason. We’ve been trained on that; 
there must be technical support if it costs 8 or 10 soles. For this purpose, a cost 
study is needed. Then each procedure has its support. That is the contribution 
we received. Another contribution is that we submitted our first draft to them, 
in order for them to analyze, evaluate and make observations on it.” 
This kind of testimonial clearly shows that the type of training provided and the topics selected 
were appropriate and have produced some degree of change in institutions. Similarly, in Ucayali 
the presence of more qualified personnel in the regional government due to PRODES programs 
was mentioned: "In our region we have very solid groups of project developers. In addition, 
we have project evaluators, such as the head of our Planning and Investment Office (OPI), 
who completed this certificate program, as did the head of the regional OPI. Having this 
fine group, we can understand that the quality of projects has improved."
Lastly, in San Martin there is a tendency to think that PRODES training sessions in the area, along 
with other activities they organized (which implies acknowledgment that there is alignment in 
the selection of the actions carried out), produced changes in organizations, as emphasized by 
the following testimonial: "At least in San Martin, goodwork is being done, and I think that is 
the result of the technical support PRODES has provided to some officials.” 
It should be emphasized that in many cases, it is acknowledged that the activities carried out by 
PRODES were not limited to training, but also included technical assistance. When this is 
recalled, the evaluation of the organization’s activities improves. 
According to testimonials gathered, technical assistance was provided with regard to citizen 
46 
participation. In any case, it is clear that technical assistance or accompaniment was provided at 
different levels. 
“- We were always given roles or commitments that we had to fulfill one way or 
another. That was your commitment. I’ll train you, and in return, you have to do 
your part. One of the impacts is the application of the results-based 
participatory budget methodology and everything that can come from it, such as 
accountability. That was PRODES’ work and it has been institutionalized." 
(District municipality of San Martín) 
The results obtained clearly show that even if the assessment is directly related to the position of 
the respondent, it can be stated that the training activities have had a positive impact on 
different geographical areas and on different levels. This means that testimonials reveal changes 
in the way they operate and their ability to meet the demands of their work both in regional and 
municipal governments. Positive comments were received from functionaries as well as elected 
officials (as in the case of assembly members). 
This reveals not only the project’s concern regarding capacity building in local and regional 
governments, but also specific contributions that are clearly differentiated from other initiatives. 
The information obtained also clearly indicates that PRODES expanded its activity beyond the 
scope of this study (this occurred with both UNICEF and PRISMA; the latter noted that by joining 
efforts, they were able to reach 14 regions). While this situation is positive, it is also another 
factor limiting the possibility of conducting a quantitative impact evaluation, as mentioned 
previously. 
That those consulted consider PRODES’ performance regarding capacity building to be good does 
not mean they have no objections or suggestions for improving the effect achieved through 
training. The most important of these is related to the need to migrate more intensively toward 
technical assistance. This is because the problems faced by public authorities and officials are 
often very complex and require follow-up, and also because there is criticism of those who think 
that a course is sufficient: 
"- Fortalecer capacidades, resolviendo casos, dejando de lado ir con el cursito o 
diplomado donde la capacidad de retención es bien poca y la gente después de 
trabajar duro, tiene que tomarse un fin de semana para entender cosas. Después 
el que enseñó se va, y la capacidad de retención o aprendizaje es pobre, el 
enfoque debe cambiar a darse asistencia técnica. Trabajar con ellos, conversar 
directamente con las autoridades. 
E:¿PRODES tuvo este enfoque? 
‐ Si un poco. Piensa en la Secretaría de Descentralización, tiene un plan de 
desarrollo de capacidades que son un grupo puntual de diplomados o cursos. 
Entiendo que políticamente tienen que atender sus necesidades, pero también 
en una oportunidad les mencioné que tenían que trabajar con ellos en el campo, 
con los consultores metidos en la zona. Tener el apoyo a nivel de universidades 
locales para que empiece a florecer el tema."
(Former conflict officer) 
Rather than a criticism of the way in which PRODES has been operating, this type of 
perception should be taken as information aimed at reinforcing aspects that 
differentiate this initiative from those existing previously. As noted before, technical 
assistance is present and must be emphasized. 
47 
One factor in favor of focusing on technical assistance or giving more importance to this 
part of the component is the possibility of ensuring that what is learned remains in the 
institution. This concern is based on high staff turnover, especially mid-level 
functionaries (although turnover is high among authorities, considering that only two 
regional presidents were re-elected), which gives rise to the need to continue with 
training. Respondents see this constant turnover as one of the major limitations of 
training. Some even mentioned that the persons trained end up as consultants that do 
not even work in the locality. Although one wants to believe that the effort invested in 
training is not lost because it ultimately enhances the capacities of those who sooner or 
later are in a public position, the fact is that the environment in which it operates limits 
the institutionalization of good practices or of improving processes in beneficiary 
entities in a sustainable manner. In addition, not everyone has the same learning 
capacity and not every functionary is willing to follow the program. Therefore, an 
official of the district municipality of Ucayali remarked, “I would suggest to PRODES 
that assistance should be more continuous and should be accompanied by 
monitoring." 
Along the same line, there is concern about the timing of training. Training during the transition 
of one administration to another is remembered as an error of PRODES-I. It is considered positive 
that this time the project is being executed well before the change of authorities. This requires 
taking into account the electoral timetable for such actions. 
Respondents believe that training on other topics or tasks that sub-national governments have 
pending should continue. Among these are everything related to administrative simplification, 
which is an important topic, considering that a lack of it hinders project execution, according to 
information obtained from the MEF respondent. In some cases, what is requested is that training 
on certain topics be extended to other levels of government or other areas. This occurs in the 
case of the TUPA, a topic that appears to be less of a concern among populations in the most 
remote areas. 
 "E: How would you evaluate the training? 
- Well, I've participated in one of the Supervisory Agencies for State Contracting 
(OSCE) through PRODES. It’s new to me, because it is not, shall we say, the area in 
which I work, but the one time I participated I found it very interesting, as you can 
see many things. I've learned many things, but now I find certain details complicated. 
However, I think it’s very important to continue with the issue of the TUPA. There are 
many municipalities in our country, in both rural areas and urban areas. The coast 
contributes more, but in our jungle and in the mountains, people are still not used to 
contributing and many of our municipalities lose a lot of income for their 
communities because they are not knowledgeable about what to charge and what not 
to charge. Work on the matter of capacity building on the topic of these 
administrative instruments, basically, should be continued.” 
(Municipality of San Martin) 
Others believe that continued training on the SNIP is needed and would be justified, given the 
large number of municipalities that have not had the opportunity to learn how to fulfill this 
institution’s requirements. Lastly, PRODES’ allies request focusing on matters that are less 
operational than those covered so far: "PRODES 2 thoroughly covered the content of the 
original plan that we made in the National Council for Decentralization with MEF and USAID. 
It appeared to be a simple topic; planning was here, projects were here and budgets were 
here. Then what PRODES 1 did very well was to improve project management, but there are 
other topics that were not handled in the same way. One of these is planning, in which there 
48 
is little interest. That would be a good framework for foreign assistance in general. We 
shouldn’t neglect the other eight quadrants that are so important.” 
An infrequent but no less important criticism has to do with the so-called “theoretical” approach 
of the training courses, referring to the development of proposals that respond to or assume ideal 
models of relationships between institutions and/or situations in the process, without taking into 
account, among other matters, the political pressure that public opinion or demands from the 
population could exert (or including information on how to handle it). This weakness could be 
eliminated with more practical learning activities, including technical assistance and/or 
increased joint work on projects geared more toward the population, such as ProParticipación. 
3.4 Main findings 
 The capacity building program has had a positive impact that is demonstrated in 
the quantitative research by the greater number of public investment projects 
approved in the regions where the intervention took place than in those where it 
did not. 
 According to the qualitative information gathered, the positive effect is due to 
providing innovative training with adequate materials. Every respondent 
acknowledged the positive features of the training. 
 Not being able to make more significant impacts in statistical terms for the 
number of projects indicator is apparently related to the turnover of personnel 
that occurs in governments, particularly at election time. Therefore, San Martin 
and Ucayali show better results than Ayacucho (where there were no re￾elections). 
 The fact that the training sessions were appreciated had to do with adequate 
selection of instructors, the development of materials, the methodology and 
every other aspect related to the logistics of the events. This indicates that it is 
necessary to continue with this planning process in the training programs. 
 The topics covered in the training sessions met the demands of regional and 
municipal governments, meaning that the process for determining them worked, 
which is vital for the demand-based approach. 
 In the opinion of participants, when training was accompanied by technical 
assistance, better results were obtained. 
 Technical assistance would be more appropriate in high-turnover contexts to the 
extent that change is institutionalized, based on comments made by those who 
had the opportunity to participate in these experiences. 
 Training requiring continuous attendance in face-to-face classes is not always 
best for senior staff. Technical assistance may be better suited to their needs. 
 Achieving good performance of regional and municipal governments is not only 
related to technical matters. Respondents indicate the importance of political 
considerations, which ultimately have to do with each authority’s style with 
regard to follow-up on processes. 
49 
 The alliance with UNICEF and introducing the human development component in 
public investment projects helped to draw attention to issues related to women 
and vulnerable populations. 
 The training of council members (male and female) received positive evaluations. 
 According to the information received, those who participated in these training 
sessions showed better performance than before. 
 TUPA training appears to be more clearly related to changes occurring in the 
institution. This is related to the matter of technical assistance, as well as the 
fact that in this case, it is easier to follow the technical recommendations. 
 When reference is made to training on procurement processes, difficulties due to 
the trend of constant changes in the legislation are mentioned, as this situation 
creates a need for continuous training. 
 District municipalities are demanding a longer intervention. It should be noted 
that they are the ones that place the most emphasis on the fact that the people 
who received training are no longer there; that is, apparently there is greater
turnover of personnel at the municipal level. 
 Training creates important opportunities for interaction by including personnel 
from different areas. 
 The pending demand continues to be diverse, which coincides with one of the 
aspects of the tailor made option (flexibility), although the problem is some 
governments’ inability to clearly establish their priorities. 
 
50 
Evaluation of the program 
characteristics 
In this final part of the document, some of the general characteristics of the program are 
analyzed in depth. The first thing to consider is that it is confined to respondents' perception of 
“foreign assistance”, and more specifically, international foreign assistance. 
There is a positive perception of international foreign assistance in the areas surveyed, 
where respondents identify it as a critical element in achieving good governance. Along 
that line, in the case of the San Martín Region, its president’s capacity to obtain 
international foreign assistance in the region was mentioned. Similar comments were 
made regarding some districts or provincial municipalities. Beyond the existence of a 
demonstrable positive correlation or lack thereof, what is important is that those 
interviewed feel that it exists. Hence, the importance of cooperation as a guiding 
element of government action: although not explicitly mentioned, it is assumed that one 
must have a good relationship with it, which implies openness to considering what it 
proposes. 
The main findings regarding international foreign assistance are shown in the following graphic. 
Graphic No. 4 
Perception of international foreign assistance 
When it is mentioned that appreciation of foreign assistance varies from one area to another, it is 
not a range of positive to negative impressions, but rather of degrees of enthusiasm about it 
regarding whether or not it is in line with the needs of the area where an intervention is taking 
place. National entities have an outlook of continuity with greater impact than smaller municipal 
governments usually envision, although the latter feel they are in great need of foreign 
The perception on
cooperation depends
on respondents’
place of residence.It
is best in Lima and
San Martin.
The perception on
cooperation is more
positive at the
regional level than at
the municipal level.  
It is better in
provincesthan in the
districts.
Cooperation is
differentiated
between that
focused  on
strengthening civil
society and that
having to do with
providing support to
authorities.
Cooperation is
differentiated
according to whether
it is directed toward
financing works or
improving services or
toward a more
institutional matter.
Cooperation is
perceived as
necessary. In many
casesit performs a
function that is
abandoned after it is
withdrawn.
There is a perception
that the capacity to
coordinate with
cooperation is an
elementthat has to
do with
governments’
capacity to do things.
Chapter 4
51 
assistance. This is probably because they have a different perspective. Central government 
authorities think of foreign assistance entities as allies (they help them to do a better job), while 
in other cases, recipients see themselves as beneficiaries (the smaller the size of the government 
receiving foreign assistance, the more they see themselves this way). 
Respondents see foreign assistance as having to do with aid to governments on three specific 
matters: financing of civil works, institutional development / improved administration and the 
strengthening of civil society. Often, governments’ greatest expectations are related to the 
financing of civil works. 
While international foreign assistance is perceived as being necessary, this does not mean it 
should not be improved. In some cases the need for better coordination between organizations 
working in the same area or with similar objectives was mentioned. In this regard, it is important 
to analyze how PRODES fits in with other international foreign assistance efforts taking place in 
the area. The answer, according to indications in previous sections herein, is that there is no 
impression of overlapping. In general terms, U.S. foreign assistance is positioned with more 
institutional issues, while other countries are more concerned with the economic environment 
(economic development programs). This does not mean that the impact of programs such as crop 
substitution is not acknowledged, but there is a feeling that it comes with a more general 
proposal, immersed in the promotion of institutionalization and democracy (a political analyst 
expressed his opinion along these lines). On the less positive side (the extent to which people see 
intervention as having to do with very specific interests), it is more closely associated with 
intervention in areas where drug trafficking is taking place. 
When referring to the limits of international foreign assistance, respondents allude to excessive 
orientation of training. With regard to this situation, they question its potential impact due to 
lack of follow-up (which has more to do with technical assistance than training), high staff 
turnover, lack of time among high-level officials (implying that training opportunities come late 
when they are already authorities, due to which developing leadership schools for those planning 
to run for office at some point was proposed). One aspect that goes beyond the focus on training 
is the heterogeneity of the areas intervened. According to comments received, the actual 
situation of the different regional and local governments in which intervention is being planned is 
not always taken into account. 
Discussion of international foreign assistance involves referring to different entities, different 
sources of financing. Basically, German, Canadian, Spanish and American foreign assistance are 
distinguished. Regarding American foreign assistance, mainly USAID and AID are mentioned. 
Although respondents noted that they did not perceive any overlapping between cooperating 
entities during the process being evaluated, the possibility of this occurring should not be 
dismissed, given that GIZ has also decided to become involved in management issues. 
52 
GráPH N° 5 
La percepción sobre la cooperación norteamericana 
When referring to U.S. foreign assistance in the area, Chemonics is mentioned (more so in the 
case of San Martin), as well as PRODES and ProParticipación. Between the last two (both linked to 
USAID) respondents see a clear difference. While the former is associated with work among 
authorities, the second one is seen as strengthening civil society, although at some point, people 
felt that PRODES also carried out actions that benefited civil society. In general terms, they 
perceived that the programs complement each other rather than overlap. 
Some respondents perceive that PRODES and ProParticipación represent not only a choice of 
different target publics, but also different ways of presenting USAID’s image, with the PRODES 
approach being a limiting factor in influencing the regulatoryagenda: 
”PRODES is a project which AID gives to a contractor; therefore it is an AID 
project. AID and this institution are the official foreign assistance of the U.S. 
government and consequently, their link has to be with Peruvian government 
entities. In a different case, foreign assistance could collaborate with civil 
society, so it would not be an AID project, but rather a group of civil society 
institutions that use other existing mechanisms in their foreign assistance. The 
Propuesta Ciudadana (Citizen Proposal) people can appear on television with a 
clearer, more political position because they are not committing AID, but rather 
a citizen proposal. " 
(Assembly of Regional Governments official) 
As indicated in the preceding point, PRODES established alliances with a number of entities. 
However, there are demands for greater coordination, which tend to come from entities related 
to citizen participation: "PRODES should seek greater articulation in its intervention. Even 
though it articulated with other USAID programs because USAID required it, with other 
projects there was insufficient articulation with parallel organizations that are working on 
similar things to improve the quality of the intervention. I think an effort was made to do 
that last year, but this aspect was neglected in the first two or three years.” (Mesa de la 
Concertación de Lucha contra la Pobreza San Martin / San Martin Round Table on the Fight 
Against Poverty) 
Work has been
done for  many
years on topics
related to
Decentralization
.
It has two
differentiated
branches:
PRODES geared
toward
authorities and
ProParticipación
toward civil
society.
Its programs are
perceived as
complementary
rather than  
overlapping.
There is a clear
differentiation
with
cooperation
from other
countries.  
There is a
positive
impression
among those that
have had contact
with it. This does
notmean they
see no roomfor
improvement.
There is a
positive
atmosphere for
its presence,
including a
demand for
continuity.
53 
Focusing on PRODES also entails discussing the demand-based approach. This concept is not 
always clearly identified by respondents, although flexibility is recognized as a feature of the 
program. For example, in the assembly interviews, it was mentioned that the demand-based 
approach is relative, because the main concerns of the cooperating parties are also present, in 
this case the war on drug trafficking: "In theory, it is the right approach. However, there is a 
catch, which is that I identify the needs of the people and, as I'm the one with the offer 
based on the needs of people, they’ll come to me because I’m the one with the offer and it’s 
free. I’d say that cooperation is a more reasonable approach. It’s an approach based on 
relative demand. Sometimes the approach is conditional on the crop substitution policy."
Suspicions aside, and taking into account the information provided by the people who were in 
charge of the program, there has been an orientation toward working with the beneficiaries 
identified, for whom the intervention was designed. Has the demand-based approach been 
positive in the achievement of the project’s objectives? There are arguments both for and against 
this approach. On the positive side, there is the ability to adapt to an extremely heterogeneous 
situation such as that of sub-national governments and even at the regional level. Another 
positive aspect is working on what the authorities are willing to prioritize, which assumes greater 
commitment on the part of governments. On the negative side, there is the possibility of losing 
sight of the program’s objectives, due to having a perspective based on the context and 
depending on the beneficiaries’ capacity to establish their needs adequately. For this reason, 
there are critical views of this process in institutions that could emulate the approach if they 
consider it appropriate: 
"The logic of the PRODES process was demand. They had defined areas or 
regions, and its analysts were those who could identify local demands and act on 
those demands. The advantage of this is that it guarantees commitments for 
joint action, but the disadvantage is that one cannot reach everyone because not 
everyone has the same demand. There are others have not established the same 
need and they are the ones in greatest need. Unfortunately, those who make the 
fewest demands are those with the greatest needs, but they lack the ability to 
make demands." (Prisma).
Regarding this point, it should be noted that improvements can be achieved in project 
formulation, but not on proposal of broader development plans whose preparation is not 
always valued in the short term due to pressure to obtain results. The validity of this 
information is evident among those who insist on following the line of providing training 
for council and assembly members first. 
The interviewee from ProParticipación said, "The advantage is that you go where you are 
needed and there you see possibilities that your intervention will be beneficial. The 
disadvantage is that if you are constantly running from one place to the next, you don’t 
establish long-term processes." 
The results obtained from quantitative analysis lead us to believe that while the demand-based 
approach led to programs that were differentiated from one region to another, this situation did 
not lead to similar results. This is not due to errors in the identification of demands, but rather to 
differences inherent to dealing with regional governments with varying degrees of 
institutionalization and development. 
PRODES is seen as a program focused on improving the capacity of regional and municipal 
governments. To a lesser degree, its involvement in the regulatory field and the development of 
tools to help governments and opinion leaders to better evaluate progress on the decentralization 
process and to evaluate the performance at different levels of government are acknowledged. 
54 
This occurs because the focus on training and technical assistance is clearer. Those who have 
worked for the PCM state that "PRODES had its role, providing support through publications 
and training that were of great help in clarifying the picture, because unfortunately changes 
of authorities always bring that sort of problems." 
The problem with training is that it must deal with the political situation in each location, an 
aspect which is not taken into account when selecting and designing interventions. In order for 
training and technical assistance to have a positive impact, minimum conditions are required on 
the following matters (according to information gathered in interviews): institutional stability, 
existence of leadership, continuity in government (which reduces turnover somewhat), 
articulated civil society, an economic development orientation, good budget execution, the 
ability to consider investment quality and foresight regarding development. These conditions 
(existing or non existing, in addition to the level attained) are the minimum conditions that 
should exist when intervention areas are selected, even more so in the case of the demand-based 
approach). Otherwise, the project will run the risk of having to deal with situations such as the 
following: "...In the case of some departments, clashes and the lack of consensus and 
dialogue hinder progress; it’s complicated....I’ve talked with Cesar Villanueva quite a bit to 
work on the matter of strengthening regional governments, since there are negative 
perceptions that spread and result in a number of problems ranging from recalls to internal 
conflicts.” (MEF official) 
This is where the need to design interventions that help balance the distance 
between one area and another arises. That situation leads many to doubt 
whether the decentralization process is something that should progress equally 
for everyone. ("Before getting concerned about how the transfer process is 
going, let’s be concerned about whether they are ready to assume new 
functions. It seems to me that transferring education to them is only going to 
increase educational differences among Peruvians, because there are great 
distances between one regional government and another.” 
(Analyst and researcher) 
The heterogeneity of contexts must be taken into account when analyzing the results for each of 
the regions evaluated: San Martin and Ucayali have re-elected administrations that differentiate 
them from Ayacucho. In turn, the dynamics in San Martin appear to be different from those of 
Ucayali, mainly because their views on corruption are markedly different. In addition to this 
heterogeneity, there is that caused by serving regions, provinces and districts. Is it appropriate to 
work on all these levels simultaneously? The answer has two dimensions. On the one hand, it is 
true that it is difficult to measure the impact obtained with the different actions carried out, 
while on the other, it is undeniable that benefits were obtained. 
PRODES’ way of working has brought about articulation (an unexpected benefits of the 
intervention), which is extremely necessary in light of the loss of spaces that should produce it 
(National Council for Decentralization/CND) and the low level of efficiency in achieving this 
objective in bodies such as the Decentralization Secretariat of the PCM: "The National Council 
for Decentralization had several problems, but a great advantage was that in that space, the 
three levels of government were trying to work toward decentralization together. What the 
Decentralization Secretariat did was concentrate all decision-making and lines of action in a 
second floor office of the PCM and end coordination among the three levels of government." 
(Assembly of Regional Governments) In light of this situation, it is important not to lose sight of 
articulation as a means of creating a climate conducive to decentralization. 
From the discussion in this section, the following findings can be indicated: 
55 
 PRODES II is a program that fills an important vacuum in terms of articulation 
among central government entities, nongovernmental organizations and the 
different levels of government. Its absence raises the need to establish who will 
assume that role. 
 PRODES II is generally better known by respondents for its capacity-building 
activities than those related to regulation, which does not mean it did not have a 
major role in this matter. It is not easy for international foreign assistance to 
influence the regulatory framework, much less to ensure that certain political 
processes take a given direction. The decentralization process is a political 
process, not just a technical process. This fact should be taken into account more 
in decisions to be made on the next part of the project. 
 The demand-based approach is controversial. The main concern is that the 
beneficiaries themselves do not know how to identify their priorities. The main 
advantage is greater commitment to take action on the part of beneficiaries. 
 To optimize the demand-based approach, an institutional assessment of the 
governments with which the project will be working is necessary. 
 There is no statistical evidence that the demand-based approach has been 
decisive in the positive results achieved by the program. 
 Authorities and technicians must be aware of the processes that public 
administration demands. Therefore, both levels must be trained. The problem is 
that training authorities is extremely complex. 
 The program’s best performance on the matter of capacity building has to do 
with methodological innovations and the quality of the activities carried out 
(including materials). 
For better achievement of the objectives, the following recommendations should be taken into 
account: 
 Strengthen alliances with all kinds of entities. This will ensure that the 
experiences gained are reproduced by other actors in other settings and 
contribute to institutionalization of the practices. 
 Coordinate with other sources of international foreign assistance to ensure that 
there are no overlapping efforts. 
 Ensure that PRODES and ProParticipacion have similar intervention areas and 
align their work better. 
 Articulation should be a more specific objective. This implies having the 
objective of fostering more explicit regional alliances and associations. 
 
56 
 Conclusions
From the data gathered over the course of this study, we submit the following 
conclusions: 
In relation to the regulatory framework of decentralization: 
 Currently, the decentralization process is going through a period of stagnation, 
one cause of which is the lack of a clear definition on the part of the central 
government about what they want to do with this process. There is no consensus 
about which direction decentralization should take, there is no strategic vision, 
or at least none has been made public and this hinders efforts at the national 
level. This situation will affect the execution of PRODES’ next phase and the 
expectations that develop around the project. 
 The progress that decentralization process has made in recent years owes itself 
primarily to the momentum coming from the country’s interior and 
fundamentally, due to the advances achieved by regional governments. They 
could become the catalysts of this process, a role which will require continued 
support and particularly more technical assistance. 
 In order to ensure that the decentralization process moves forward with the 
regulatory framework, policy spokespersons are needed. It is essential that 
regional and local governments have a greater capacity to develop proposals and 
exercise influence and leadership in order to have bearing on decision making. 
 The materials developed by PRODES are appreciated for their quality. However, 
these need be given more visibility in order to become a real benchmark or a 
common starting point in other areas where the project currently has no 
presence. 
 The diagnosis performed using the Decentralization Map and the Decentralization 
Compass is still valid and the aspects identified as needing more work to be able 
to move forward with the decentralization process are still pending. 
 The task that still needs to be carried out is building a consensus, mainly in 
relation to priority actions. Insofar as this occurs progress can be made in 
preparing a regulatory framework that fundamentally resolves the definition of 
the powers and jurisdiction of each level of government and the ways in which 
they should interact and coordinate. 
 The Congress of the Republic of Peru lacks the resources to replicate the 
processes involved in producing the Decentralization Compass, which they 
consider necessary in order to update the document. The document itself is 
Chapter 5
57 
perceived as too heavily focused on the PRODES II project areas of intervention 
and Congress may open up the debate at the national level. 
 The technical assistance used to prepare legislation has been an effective way to 
overcome the inherent limitations of being part of international foreign 
assistance in helping to enhance the regulatory framework. 
 The creation of spaces for coordination between the different levels of 
government was one of PRODES correct choices for minimizing social conflict. 
 Social conflicts represent a complex set of problems where, as with 
decentralization, the authorities and responsibilities of each level of government 
have not been defined. Confrontational situations are created due to the lack of 
clarity in these roles, which makes conflict resolution and coordination efforts 
difficult. 
 Politics carries a good deal of weight when trying to deal with social conflicts. 
Because of this situation, an approach that is exclusively technical, an approach 
primarily dedicated to building capacities through the sharing of methods that 
are useful in social conflict management will rarely succeed. 
 There is no consensus about what the causes of these social conflicts are, a 
situation that hinders the application of the demand based approach. This, 
insofar as government authorities have not necessarily been able to identify what 
their needs actually are. 
 Adequate conflict prevention/resolution management requires unique approaches 
for each region or township; since there are underlying problems that need to be 
addressed, similar to what occurred with the artisanal and small-scale mining 
law. 
About the gender approach 
 We have observed positive impacts on the lives or professional development of 
the participants in program activities and those in their environment. 
 The activities carried out by PRODES have allowed initiatives to be carried out 
that are aimed at reducing the gaps between men and women and laws or actions 
aimed, for example, at giving women more visibility whether as government 
employees or as beneficiaries of concrete actions on the part of the government. 
 The training workshops on the subject of budgeting have helped to raise aspects 
or issues that are more valued by women and were not being taken into account 
by the authorities. This refers to aspects related to the development of projects 
related to human development or related to childhood, which were submitted by 
those who participated in the training programs for councilmen and council 
women, assembly men and assembly women, also participated in drafting 
ordinances that included the gender approach. 
 Those who participate in the council men/women, participate in the 
development of ordinances that included the gender focus. 
58 
 Gender topics were included in all activities directed toward both men and 
women and not only those organized exclusively for women. 
About capacity building 
 The PRODES intervention generated a positive impact with its presentation of 
projects (in a quantitative format) and actions have been identified which have 
shown governments to be more efficient as a result of activities implemented by 
PRODES. 
 Technical assistance would be more appropriate in contexts of high rotation, 
insofar that change is institutionalized based on what has been expressed by 
persons who had the opportunity to take part in these experiences. 
 To become more efficient, capacity building needs to incorporate both training 
and technical assistance, with more emphasis on the latter. Technical assistance 
is what enables us to be more likely to be successful in having individual 
knowledge transformed into institutional change. 
 The PRODES approach strengths include that it forms alliances, develops 
innovative methods, designs quality materials and responds to the needs defined 
by the beneficiaries. 
 The PRODES model weaknesses include the heterogeneity of the intervention, the 
fact that the users are not always clearly aware of what their needs, the 
difficulties in attracting the most senior public officials to participate in the 
training along with the high staff turnover rates at the levels of government 
where we have been working. 
 Capacity building is still in demand at different levels of government. However, 
we need to be more diligent when identifying if what they really need are the 
abilities to apply established procedures or if there are other more fundamental 
problems (the lack of the ability of political leaders to make correct decisions). 
 Training assembly persons and council members (male and female) is a way to 
assist political leaders to develop a clearer vision about their responsibilities and 
how to help satisfy the demands of the population. 
 That the training programs for assemblymen/ assembly women and councilmen 
/council women incorporate a vision of gender that should be accentuated in 
order to generate a greater number of approved regulations on this issue. 
Program characteristics 
 The demand based approach has been a source of differentiation for this project. 
However, there is no evidence to suggest that the program’s achievements owe 
themselves to this aspect, or that this approach is the only way to gain 
beneficiary trust and commitment. 
59 
 The demand based approach has lead to a dispersion and heterogeneity that does 
not allow comparison of the results obtained in one place or another 
 The application of the demand based approach should include an in depth 
analysis of the institutional situation of the municipalities subjects of program 
intervention. 
 The work carried out with the regional governments, provincial municipalities 
and district municipalities has resulted in the benefit of creating spaces for 
interaction and coordination. This work method can be counterproductive in 
terms of the amount of impact it generates because it leads to the dispersion of 
activities and resources. 
 PRODES has carried out its program in locations where other USAID sponsored 
programs and international foreign assistance agencies operate, without 
overlapping efforts. There was a case with a US foreign assistance agency where 
efforts were coordinated to complement each other, between Pro Participation 
(aimed at the population) and Pro Decentralization (aimed at the authorities). 
 PRODES continues the trajectory of US foreign assistance, which is strongly 
devoted to the aspect of institutional strengthening. 
 
60 
 Lessons learned
In relation to the PRODES II Project, we list the following lessons learned andrecommendations: 
 While as an agency of international foreign assistance, USAID does not take a 
leadership role in promoting the institutionalization of intergovernmental 
coordination as a part of the decentralization process, it is recommended that it 
continue to offer neutral spaces of interaction between the key players in the 
National Government (Executive and Legislative Branches), the Regional 
Governments and Local Governments, with a view toward paving the way for the 
formation of networks and the development of synergies that could potentially 
help to build a new consensus as to which direction the decentralization process 
and its applicable legislation should take. Any space that is conducive to the 
development of intergovernmental relations based on trust and cooperation 
between the actors of these three levels of government can contribute to 
overcome this process’s present stagnation and move forward with developing 
the key policies needed for successful decentralization. 
 Although since its inception decentralization has been conceived in legislation as 
a top down type of process, it would be meaningful to reinforce USAID 
interventions by supporting the development of consensus and articulating 
existing proposals on the guidelines of the decentralization process at the sub 
national levels of government. The processes of horizontal coordination should be 
enhanced since this can lead to a Bottom- Up type of dynamic or the formation of 
coalitions dedicated to encouraging policies that contribute to resolving the 
problems that currently have the process at a standstill. Along this line of 
thought, a deeper alliance should be achieved with the Assembly of Regional 
Governments, as well as with the entities that assemble the municipalities. 
 With regard to the Decentralization Compass that was recently transferred to the 
Congressional Committee on Decentralization, it is advisable that it be monitored 
to ensure its continuity and particularly to ensure that it remains an instrument 
used to gather and articulate the points of view voiced by the different national 
actors. 
 As for regionalization, we recommend continuing to work with initiatives related 
to organizing town associations and with the Interregional Coordination 
Committee as processes which seek the development of a dynamic and consensus 
that enables the process to move ahead. 
 In relation to regulatory reforms which require political will and a high degree of 
consensus, it is recommended that USAID continue to maintain a low profile, 
offering technical assistance and foreign assistance as a function of what key 
decision makers require and opportunities the context offers. 
Capítulo 6
61 
 It is recommended that emphasis be place on working through strategic alliances, 
given that they allow one to join forces and obtain resources, and they generate 
a broader impact from the activities carried out it is likely that activities 
continue to replicate themselves beyond the PRODES PROJECT scope of 
intervention. 
 The beneficiary entities perceive and appreciate project activities more 
positively when there is greater presence of technical assistance. More emphasis 
must be placed on technical assistance. Migrating more intensively toward 
technical assistance must be accompanied by continuous supervision. 
 The electoral calendar should be taken into account when scheduling program 
activities, particularly training activities. This also relates to the need to 
evaluate the existing level of institutionalism in the area of intervention. 
 In regard to the future of capacity building activities, we need to consider the 
following: 
a. Training strategies need to be established that consider the high government staff 
turnover that takes place following an electoral process. 
b. When training high level government officials, we need to provide them with 
elements that help them understand the importance of building and maintaining a 
career in public service, and to work as career officers. 
c. Mechanisms should be found to ensure that training becomes more institutionalized 
and beneficial to the governments selected for program participation. One 
possibility is to establish that government officials provide their commitment to 
remain with the entity and with those who receive training identifying the 
expected results or goals. In this manner they would utilize foreign assistance 
agency and government resources more effectively. 
d. More emphasis must be placed on encouraging the adoption of good practices, 
which means being able to clearly identify which regional, provincial or district 
governments are implementing good practices and with what results and then to 
findthe mechanisms to provide visibility and replicate them. 
e. We must continue to ensure that the training strategy considers the quality of the 
professionals, methods and materials. This factor was the reason that motivated 
everyone who received the training to recognize the relevance of the courses, the 
appropriateness of the subject matter and above all the existence of results that 
are a consequence the training itself. 
 One relevant challenge consists of improving the availability of data in order to 
conduct a rigorous assessment at the time of a program’s inception. In this 
regard, the barometer should have the information on the number of PIPs 
prepared in its data base so that we can determine the ratio of the No. of PIPs 
approved/PIPs prepared, which is better indicator of efficiency. 
 In relation to the above, while we have access to figures regarding the total 
amounts of approved Public Investment Projects at the end of each year, it is 
62 
worthwhile to have available data on the amounts involved in each investment 
project or other relevant indicator for each and every project in every 
government entity over the course of the entire sampling period. 
 In the future we should seek a stronger relationship between the program design, 
its implementation and data collection, with a view toward increasing the 
possibility of performing our assessment with rigorous methods. Data collection 
should be consistent with the proposed assessment method and defined from the 
beginning of the program. In this regard: i) we suggest reviewing a reduction of 
the heterogeneity in the training, and ii) keeping more detailed and 
computerized records on the parties who receive training and of participating 
government entities. 
 We suggest that certain aspects of the demand based approach be revised. 
Beyond its advantages the approach can differentiate its effectiveness depending 
on the relative capabilities of governments to correctly identify their needs. In 
this sense, the demand based approach employed by the program could be 
supplemented with an evaluation of the corresponding specialist so that training 
content is tailored to the needs of the regional and municipal governments and 
not solely because these governments request it. This point becomes more 
relevant in the case of a variable difficult to model such as the Execution of the 
investment budget. 
 Given that there are signs of a positive effect in the number of approved public 
investment projects, the program should make recommendations to the 
governments about the importance the role that job stability of government staff 
and employees who work in these relevant areas represents, or ensure that the 
capacities acquired as a result of the program are not lost due to government 
officials and staff turnover. 
 In the case of training workshops on the subject of budget execution, the 
program should meet with those involved (government entities, experts, etc) to 
look at performing a diagnosis based on the determining factors of budget 
execution, such as the professional profiles of human resources, staff stability 
and turnover, compensation policies, the quality of procurement and logistics 
processes, among others. The analysis should seek to distinguish whether these 
aspects are more relevant and if this is the case, determine whether it is 
economically viable to multiply training efforts in terms of quality or quantity. 
 The gender approach should be maintained and even enhanced in order to 
achieve that the projects, ordinances and regulatory designs always include these 
principles. The target objective should always be to increase the number of 
approvals. 
 
63 
 Theoretical framework of the study
The proposed evaluation design comprises two approaches: one is quantitative and the other 
qualitative. Below we present a methodological description of the implementation of both 
phases. 
1.1. QUANTITATIVE STAGE 
This section will provide only a few ideas that summarize many theoretical and empirical 
developments. In principle, we will employ two approaches to evaluate the project: the 
conventional impact assessment approach (quasi-experimental) and different versions of the 
panel data model. 
If there is insufficient data to conduct a rigorous impact assessment, we will make our best effort 
to construct simple correlations and descriptive statistics of the primary outcome indicators, 
taking into account the unique characteristics of the sub national governments covered by the 
project. 
1.1.1 Quasi-experimental approach 
In order to know the impact or benefits of a program on its beneficiary population, one must 
identify the cause and effect relationships between the components (goods and services) that the 
program produces and the expected outcomes or variables of interest upon which the program’s 
benefits are defined. To achieve this, it is essential to isolate the observed benefits (specifically 
the intermediate and final outcomes), all those impacts on the beneficiary population derived 
from factors outside the program 12 , and which would have occurred anyway without the 
intervention, as a consequence of the normal course of events in the environment or the actions 
of other public or private programs. 
Therefore, to quantify the program’s impact, it does not suffice to compare the “before and 
after” status of program beneficiaries, but instead, respond as accurately as possible to the 
question: What would the situation of these beneficiaries be today if they had not participated 
in the program? This is called a counterfactual conditional situation. In methodological terms, 
the above involves comparing the program’s final outcomes (and the intermediate results as 
applicable) on the beneficiaries relative to a group of non beneficiaries with similar 
characteristics (called “control group”), and in so doing, isolate the effects of other factors. 
A control group is basically a segment of non beneficiaries who are part of the target population, 
i.e., a group that shares the social, economic, educational, etc. characteristics of the beneficiary 
 
12 To be precise it is worthwhile to isolate the program’s causal effect only from those elements that are correlated with
the program assignment (factors that are unique to the group of beneficiaries or related to the level or magnitude of the
benefit provided). Those elements which affect the final outcome but which have the same effect on the beneficiaries
and about whom there is no relation to the benefit do not need to be included in the measurement of the impact.      
Appendix 1
64 
group. The only difference between the individuals in the control group and the eligible 
beneficiaries should be their participation in the program. 
The evaluation of the ProDecentralization project proposes a unique challenge: the treatment 
units are sub national (regional, provincial and local) government entities, and so it is not a 
trivial matter to find units that are strictly comparable on observable and unobservable aspects. 
For the same reason, to the extent this is possible, we shall probe for data that enables us to at 
least construct alternative methods to control the most diverse set of characteristics and the 
differences between the treatment units and the control units and to construct the best 
counterfactual scenario. 
In summary, an impact study requires prior knowledge or assumptions about the process that 
generates the data and enables us to distinguish causality of association. The statistical methods 
for estimating impact parameters are selected according to each program’s characteristics: 
objectives, design (benefits, eligible population, mechanism of assignment to treatment, etc), 
secondary data or those that are possible to obtain through some method of (primary) data 
collection. 
As a minimum, the data contains the characteristics of the individuals (or other prominent unit of 
analysis), variables of the intermediate and final outcomes, other variables capable of 
influencing the outcomes and those originating from surveys and administrative records). Below, 
we provide notes about some relevant topics 
a) Causality and the problem of identification 
The impact assessment of a given program requires making an estimate of the causal effect, 
attributable exclusively to participation in the program (treatment D) in outcome variables (Y). 
The key problem is the association between variables D and Y, which can be due to: 
‐ Effect of D on Y (causality) 
‐ Effect of Y on D (inverse causality) 
‐ Effect of a third variable X on Y and D. (“X’s effect becomes confused with D’s effect...”) 
For example, one can ask him or herself if the training or reinforcement activities have achieved 
any effects on the execution of the budget or on the increased level of tax revenue collection of 
a local government. In this case, it would not be enough to compare the results of districts that 
received training or technical assistance and the others if for example the best results are more 
common in the regions or districts that have greater resources (technological, human, etc). In 
this case, the effect of training or technical assistance becomes confused with the effect 
generated by other inputs: if other inputs have a positive effect on achievements, we will 
observe a positive association between some districts that were received treatment and others, 
even when there cannot be any direct effect from the ProDecentralization Project. 
The program’s impact corresponds to the difference in the outcome variable (Y) recorded by 
individual i (or the sub national government) with or without the program’s intervention. 
  YY 01 iii (1) 
Satisfying the condition of independence requires the random selection of the individuals who 
benefit from the program, known as an experimental or random design. As previously stated, 
the ProDecentralization project requires a non experimental evaluation approach as long as the 
data is available. 
65 
 10   /, XDYY
b) Quasi experiments 
Matching Method 
For various reasons, a randomized control group is often unobtainable, so this methodology will 
attempt to recreate the conditions of an experiment when it is not available. By aligning the 
distribution of the characteristics of the untreated population and the beneficiary population, the 
method replicates a feature of the random (experimental) data (Blundell and Costa Dias 2002; 
Heckman, Ichimura and Todd 1997) 
The “Matching” method, attempts to estimate what the outcome of the treatment group would 
be if it had not participated in the program, starting with the pairing of each participant from the 
sample with members of the non treatment group. Once each beneficiary is paired with 
his/her/their respective control group member(s), the only difference between both groups is 
their participation in the Program. 
The quasi-experimental groups have a problem in common that arises in practice for the majority 
of programs. It is very difficult to obtain an identical control subject like that obtained in 
randomization. The Matching method rests on the assumption that the assignment to treatment 
is independent of the potential outcomes; once it is controlled by a set of observable 
characteristics (called “confounders”), it should satisfy the condition of conditional 
Independence: 
The main problem with matching estimation (in cross sections) is that if there are unobservable 
elements between the treatment and control groups that affect the final outcome, the indicator 
measuring the impact of the program will be biased.
Differences in differences (double difference) 
One condition necessary when developing a quasi-experimental design is the existence of a data 
base (or several bases that enable us to form a common database of beneficiaries and non 
beneficiaries that has information for both groups about program eligibility criteria (for example: 
socio economic status, population size, average level of education in the sub national 
government, geographic location or whatever other characteristic that enable the measurement 
of the final outcomes or expected impacts of the intervention (for example, achievements in tax 
revenue collection). 
Once the control and treatment groups are formed and having obtained the observations of the 
outcome variables and the variables that characterize the individual and his/her environment in 
two moments in time (before and after the program’s intervention), the impact of the program 
will be quantified through econometric estimation in the following model, where the estimator of 
parameter α (ˆ ) corresponds to the impact of the program: 
Y i3i21it itii *X)Pe*TR(PeTR it        (6) 
Where, 
Yit = outcome variable for individual 1 en period t. 
TR i = variable equal to 1 is individual i belongs to the treatment group and 0 if he or she belongs 
to the control group. 
66 
Pe i = variable equal to 1 if the data from the outcome variable for individual i belongs to the 
post program period and 0 if not. 
Xit = set of variables which characterize the individual and his/her environment. 
it  = random error term 
The estimation of the model of equation (6) is known as the double difference method. If 
outcome data is available for both groups (treatment and control) before and after Program 
implementation, it is possible to compare the differential evolution of such results and obtain a 
credible assessment, provided that the differences in unobservable aspects have remained 
constant over time; that is, if the average outcomes of the beneficiaries and control group 
members had followed a parallel path in the absence of the treatment, which is not always true 
(Abadie 2004).
Instrumental variables and regression discontinuity 
The bias from unobservable characteristics is a problem that leads to developing other methods 
such as the instrumental variables method. This bias is derived from the existence of factors that 
affect participation in program (Pi) and the researcher is unaware, for example of the disposition 
to participate in the training program on the part of officials or the more competent public 
employees. In statistical terms this implies that in the equation: 
Yi  21 XTR ii
*   i (7) 
The error term i contains unspecified variables in matrix Xi which affect participation in 
program (Pi). 
To correct this bias you need to use the estimated value of P, according to the formula 7, but 
including a variable Z that determines participation in the program, but has no effect on the 
results. That is, a variable that is not in X and is not correlated with the error term. This variable 
is called instrumental variable and its effect is that it allows Pi to vary independently, 
eliminating the cause of the bias (Angrist et al., 1996). 
Finally, another method that is widespread internationally is regression discontinuity, which, in a 
quasi-experimental context, seeks to use the discontinuities and delays of program 
implementation. For example, a cutoff score could be established for determining eligibility that 
can be used to obtain estimates in the neighborhood that can serve to make those that are found 
around the threshold comparable, beneficiaries and non-beneficiaries. The method has objectives 
similar to those of instrumental variables, because it uses an exogenous variable that has a strong 
correlation with participation, although it is dissimilar (Imbens and Lemieux, 2007; Khandker et 
al. 2010). 
1.1.2 Panel data model approach 
In the event it becomes too complex to construct a counterfactual scenario for each beneficiary 
and if data permits, one can put together a panel data model for the impact assessment, 
following Angrist and Krueger (1998). According Benavente (2004), a set of panel or longitudinal 
data is one that follows individuals (companies, governments, or countries) over time and 
consequently provides multiple observations for each individual. In this manner, the panel data 
combines data in a time series and a cross section. 
This model assumes a specification such as the following: 
67 
Y  iit  it XTR   itit
ݐ݅ݒ	 ൅ 	μi 	ൌ ݐ݅ݑ :Where
Note that μi denotes specific unobservable individual effects that are time-invariant. These 
account for any type of individual effect not included in the regression. In regard to the term
vit, this refers to the so called idiosyncratic effects. 
Panel data models Fixed Effects: This model assumes that the individual effects μi are 
deterministic parameters should be found. This would be the case if N represents the 
“population” such as a series of sub national governments and the inferences are only related to 
the N available. Thus, the inference is conditional on the N individuals and it is not making an 
inference for other sets. 
Panel model with Random Effects: Este model assumes that the individual effects are not 
deterministic but that each one is a random variable. This is a reasonable assumption in the 
event that the data were a genuine sample used to make an inference about the population as a 
whole. 
Hausman Test: When there are doubts about the most suitable model, we use the Hausman test, 
which simply asks if there are significant differences between the estimators for Fixed Effects 
and Random Effects. 
Dynamic panel models: These models enable us to estimate ratios of a dynamic nature and 
which is impossible to estimate using cross sectional data. For example if the tax revenue 
collections of a sub national government depends on revenue collections from the previous year. 
1.1.3 Variables to use, potential sources of information and constraints 
In principal, it is essential to have indicators that can demonstrate whether a project is effective 
or not, both for the quasi-experimental approach and the panel data models. To the extent 
practicable, we will attempt to use the two above mentioned approaches depending on the 
variability and availability of the data (for example, the frequency of the series for the panel 
data models or the availability of variables to control the differences between the different 
government entities). As variables to explain (endogenous variables Y) we will use the impact 
variables proposed in the document PERU PRODECENTRALIZATION PROJECT: ACTIVITY AND 
PERFORMANCE PLAN MAY 2009. 
It must be noted that it will make a difference when measuring the impacts between the 
government units that received technical assistance and those that received training. 
The most important results are: 
1. Increase in the percentage of public investment projects declared as "viable", 
according to the SNIP. In particular, this is the average percentage by which the sub 
national treatment group exceeds the number of public investment projects declared as 
viable. 
2. Increased budget execution per government entity. This is the average at which the 
sub-national governments in the treatment group exceeds the percentage of execution of 
public investment in the non treatment group (control group). 
3. Improved administrative efficiency and simplified registration and licensing 
procedures for new companies. This refers to the percentage of sub-national 
governments from the treatment group that has adopted the unified (simplified) 
administrative procedures TUPA. The units that received technical assistance will be 
distinguished from those who received training.
68 
4. Improved tax collection. This is measured as the percentage of municipalities in the 
treatment group who have increased their tax revenues by at least 10% above those of 
the control group. As in the previous case, those units that received technical assistance 
will be distinguished from those who received training. 
In addition, we will attempt to conduct the evaluation with the mandatory indicators proposed by 
USAID. In any case, ideally we would want to have constant variables for specifying Y. 
Another critical aspect in the panel data estimation method is that one can build data based on 
the TR intervention for every geographic unit examined and for every year, ideally from the 
period prior to the project startup year. In this sense, having data available on the budget 
executed by the project, or the number of public workers or government officials trained would 
allow to take advantage of an eventual graduated progression of the treatment "dose" for each 
geographic unit. Thus, one could make an approximate estimate of the "marginal effect of one 
Peruvian Nuevo Sol spent by the program” or "one additional trained staff member." 
In relation to the above, the estimate would only be one level below the project impact for an 
outcome Y, because an intervention that does not reflect the diverse set of activities executed 
by the project in building the capacities of the sub national governments would only be 
quantified as TR. 
According to the methodology, thoroughness is required in choosing control groups, as they are 
being selected from among the sub national governments who are not in the treatment group. 
When applying the quasi-experimental method, it is necessary to monitor the effects of variables 
such as the ones mentioned below: 
‐ Population
‐ Average level of education of the population
‐ Proxy variables of socioeconomic indicators
‐ Demographics: age groups, gender, etc..
‐ Availability of basic infrastructure
‐ Number of employees working in the public entity
‐ Budget size  
‐ Quality of human resources, among others
‐ Programs pertaining to other institutions and to the National Government
In the case of panel models, if the data allows it, a series of each indicator for each sub national 
government will be constructed that will make it possible to apply the methodology. 
In principle, the main sources of information for the quantitative part are listed below: 
‐ Data provided by officials from each one of the sub‐national governments.
‐ Barometer prepared by Peru USAID / ProDecentralización.
‐ Project baseline data gathered by the Institute of Peruvian Studies (IEP) to develop the
barometer.
‐ Data from the Economic Transparency Portal of the Ministry of Economy and Finance
(MEF), where both the revenues and expenditures of each regional and district
69 
government from 2009 to date are provided along with the progress completion of budget
execution and the investment projects of the regional governments.
‐ Data from the Human Development Index.
‐ Outlook forecasts from the national census.
‐ National Household Survey
In total, there are 192 units to observe adding the district, provincial and regional governments 
of Ayacucho, San Martin and Ucayali. The breakdown of the analysis will be done according to the 
availability of data, since it is not only essential to have the data from every target variable (Y), 
but also to have a series of variables to control the differences in each region (the control 
variables). If it is not possible to accomplish this at the district level, we will begin moving up to 
the provincial level and so on, paying careful attention to the thoroughness of the study process. 
To accomplish this it will be essential to use the data from the evaluation midpoint, as this will 
provide us with an idea of how the project was performing as of that moment. Similarly, we will 
need to evaluate whether or not there was a significant changes in the evaluation policies of 
USAID or ProDes, as this could lead to biased estimates. 
It is important take into account where the intervention applied the demand based approach 
when comparing the units, since statistically this factor can lead to biased estimates. This 
approach brings selection bias in unobservables, as determined by the needs of each unit 
intervened. However, this has been done at the regional level, so the units within each of them 
have been intervened in the same manner. Methodologically we should also eliminate bias 
related to some other intervention with similar objectives in these areas, whether or not they are 
sponsored by USAID or international organizations. Finally, it will be important to review the 
years where interventions began in each unit. For example, intervention at the provincial level 
took place the first year, while the intervention for district governments has been in phases.