International Public Management Journal

ISSN: 1096-7494 (Print) 1559-3169 (Online) Journal homepage: https://www.tandfonline.com/loi/upmj20

The Use of Performance Information Under

Performance Budgeting Systems: A Comparative
Analysis of Chilean Line Ministries

JUAN PABLO MARTÍNEZ GUZMÁN

To cite this article: JUAN PABLO MARTÍNEZ GUZMÁN (2020) The Use of Performance
Information Under Performance Budgeting Systems: A Comparative Analysis of
Chilean Line Ministries, International Public Management Journal, 23:4, 591-610, DOI:
10.1080/10967494.2018.1486344

To link to this article: https://doi.org/10.1080/10967494.2018.1486344

Published online: 20 Sep 2018.

Submit your article to this journal

Article views: 326

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=upmj20
 International
Public
Management
Journal

THE USE OF PERFORMANCE INFORMATION UNDER

PERFORMANCE BUDGETING SYSTEMS: A
COMPARATIVE ANALYSIS OF CHILEAN
LINE MINISTRIES


JUAN PABLO MART
INEZ GUZMAN
UNIVERSITY OF MARYLAND

ABSTRACT: Numerous studies suggest that performance budgeting has limited impact
on budget aggregates and on legislative decision making. For that reason, a growing
number of scholars now argue for increased attention on how performance budgeting
reforms influence decision-making processes within line ministries and during budget
execution. In this study, we compare the cases of four line ministries from a government
with a longstanding tradition of performance budgeting. We find that the use of
performance information is marginal during budget formulation and approval. In con-
trast, the evidence shows that officials at line ministries use performance information
extensively during budget execution. The use of performance information differs between
our four cases and those differences are explained by a set of variables that include
executive leadership involvement, informational system capacity, and intrinsic organiza-
tional factors such as the size of the organizations and the homogeneity of their outputs.

INTRODUCTION
For more than a quarter of a century, many governments from all regions in
the world have introduced performance budgeting reforms. These reforms aim at
generating performance information to enhance budgetary decision-making processes.
The final hope is that, by considering performance information when allocating and
using public resources, governments will become more efficient and effective.
Despite its widespread use around the globe, the benefits of performance
budgeting are not always evident. For instance, several studies have found little or
no linkage between performance information and budget aggregates, casting doubt

International Public Management Journal, Vol 23(4), pages 591–610 Copyright 2018 Taylor & Francis Group, LLC
DOI: 10.1080/10967494.2018.1486344 ISSN: 1096-7494 print/1559-3169 online
592 International Public Management Journal Vol. 23, No. 4, 2020

on the impact of performance information during budget allocation (Gilmour and

Lewis 2005; Ryu 2013; Rhee 2014). In a related line of research, there is vast
evidence that legislators tend to disregard performance information when
negotiating and approving the budget (Dean 1986; Berry, Brower, and Flowers
2000; Cheung 2006).
The fact that performance information does not affect budget aggregates nor
influence the legislature does not mean that performance budgeting reforms are
not having a positive impact elsewhere. First, on its relationship with budget aggre-
gates, it is reasonable for performance information not to have any noticeable
effect because there is no straightforward relation between performance informa-
tion and resource allocation. For example, a program with deficient performance
might require more, instead of less, funding (Moynihan 2006; Shah and Shen 2007;
Joyce 2011).
Second, given the fact that legislators tend to overlook performance informa-
tion, some scholars now argue that performance budgeting should place more
attention on decision-making processes within ministries and during budget
execution (Wang 1999; Joyce 2003, Melkers and Willoughby 2005, Ho 2011). In
fact, the tendency to overlook decision-making processes during budget execution
has not only been an issue in the performance budgeting literature, but in the
general field of public budgeting (Thompson and Zumeta 1981; Pitsvada 1983;
Hackbart and Ramsey 1999; Dougherty, Klase, and Song 2003; Cain, Choudhury,
and Clingermayer 2004). This tendency is partly explained by the fact that
decision-making processes during budget execution are harder to observe, as “the
ordinary routines of budget execution are usually carried out far from the spotlight
of public attention and the day to day crises that make up much of the daily lives
of account administrators and program managers are slowly buried under an
accumulating mass of detail as the fiscal year rolls on. Even players within an
agency see little of this budget execution process as it unfolds, unless it affects their
own program” (McCaffery and Mutty 1999:234).
These and other contemporary tendencies in performance budgeting have led to
its re-conceptualization into performance budget management (PBM). Through the
lens of PBM, performance budgeting reforms take a broader scope that goes
beyond budget-allocation decisions, sometimes overlapping with performance
management, and consider “the full budget cycle, from preparation, appropriation,
and execution to evaluation, as a platform to achieve various managerial, budget-
ary, and policy-making purposes” (Ho 2018:5).
In this study, we focus exclusively on decision-making processes within line
ministries. We place special attention on budget execution, which is the stage of
the budget cycle when line ministries have a predominant role. Our analysis consid-
ers the main dichotomy from the budget execution literature: the balance between
tight controls to ensure administrative compliance with original budget goals and
the budgetary and managerial flexibility to allow line ministries to reach those
goals (Schick 1964; Pitsvada 1983; McCaffery and Mutty 1999; Barzelay and
Thompson 2006). In addition, our approach to investigating the use of
Performance Information in Chilean Ministries 593

performance information is not through statistical tests of budget aggregates, but

through in-depth observation of decision-making processes. In that vein, the objec-
tives of this study are to: (1) understand the types of decisions, particularly during
budget execution, for which line ministry officials use performance information; (2)
compare line ministries within the same government to discover potential differen-
ces in the way that performance information is used; and (3) analyze the factors
that explain such differences.
This study focuses on the government of Chile, which is a compelling case for
this study because it has a highly reputable performance budgeting system. For
example, two versions of a survey by the Inter-American Development Bank found
that Chile has the most advanced performance budgeting system in Latin America
and the Caribbean (Garc
ıa L
opez and Garc
ıa Moreno 2010; Mak
on and Varea
2015). The finding is consistent with that from a study by the Organization for
Economic Co-Operation and Development, which shows that Chile has one of the
most developed performance budgeting systems among its member countries
(OECD 2013). To this author’s knowledge, there is only one study that directly
asked Chilean line ministries about their use of the performance information
(Zaltsman 2009). However, the emphasis of that study is different, as it focuses
mostly on budget allocation and budget aggregates.
Chile has institutionalized a performance budgeting system that generates diverse
types of performance information through its many components. Most of those
components, as well as most stages of the budgetary cycle, are dominated by Chile’s
central budget office (DIPRES). The main components of Chile’s system are:
ex-ante evaluation of investment projects; ex-post evaluation applied mostly at the
program and institutional levels; performance indicators for each institution, which
are also the basis for a merit-pay system; and ex-ante evaluation of new or redesigned
programs. In this study, we cover the performance information generated by all of
those components, as well as the one produced independently by each line ministry.

DATA AND METHODS

To compare line ministries, we adapt a research framework that was built from
a systematic review of performance budgeting reforms in over 20 countries around
the world (Mart
ınez Guzm
an n.d.). We focus on the 23 variables of that frame-
work that relate to the technical aspects of the performance budgeting systems and
to the organizational and individual factors within each line ministry (see Table 1).

Case Selection
We compare four cases, where each case is a single line ministry of the government
of Chile. Case selection was a challenging task, since we do not have ex-ante values
for our dependent variable (the use of performance information), nor for most of the
independent variables listed in Table 1. One problem of selecting cases without prior
knowledge of the dependent variable is the potential for an undetermined result; i.e.,
594 International Public Management Journal Vol. 23, No. 4, 2020

TABLE 1
Research Framework

Technical aspects of performance budgeting

Quality of performance information from centralized systems
Stakeholder responsibilities, accountability, and involvement
Availability of performance information from centralized systems
Integration with strategic planning
Characteristics of the budgetary processes (e.g., timelines, earmarks, aggregation level)
Informational system capacity to collect, store, and access performance information
Objectives and priorities of the system
Costs of processes and availability of resources related to performance-based reforms (time
and money)
Top-down or bottom-up approach to performance budgeting
Provision of training and methodological instructions
Quality of performance information from internal systems
Availability of cost information
Linkages between budget lines and performance measures
Managerial flexibility
Availability of performance information from internal systems
User-friendliness of the system
Legal framework of performance-based reforms
Organizational & individual factors
Internal leadership support for performance-based reforms
Staff buy-in for performance-based reforms
Staff capacity for performance-based processes
Characteristics of the goods and services provided
Organizational and/or program structure
Size of the organization

Note: Adapted from (Mart
ınez Guzm

an 2017).

finding the same value for the dependent variable despite variation in the explanatory
variables (Gisselquist 2014). We address this challenge by focusing on one of the few
independent variables for which we have ex-ante information and which is expected
to be relevant. That variable is the size of the line ministry.
The size of the organization has been predicted to be an important factor in
decision-making processes. For example, in his seminal work, Downs (1966) argues
that larger bureaus have weaker coordination among members and present a
harder environment for top officials to maintain control of their subordinates, a
situation that some large bureaus might be able to offset through increased prod-
uctivity from economies of scale. The situation described by Downs suggests that
the challenge of collecting pertinent and timely performance information is often
greater in larger bureaus.
Performance Information in Chilean Ministries 595

The most straightforward way of comparing the size of line ministries is through
the amount of resources in their budgets. An alternative is to use the size of their
labor force, but that measure is biased towards labor-intensive sectors. The distri-
bution of budgetary resources among Chilean line ministries does not result in a
clear differentiation between large and small organizations; instead, it shows some
very large and some very small line ministries, coupled with a diverse group of
middle-sized ones. To separate line ministries into more precise groups, we ana-
lyzed them using k-means clustering methods. Figure 1 presents Chile’s line minis-
tries, sorted from the largest to the smallest budget and divided into four clusters:
very large, large, small, and very small line ministries. This study analyses a single
line ministry from each of the four clusters. Those line ministries, ordered by their
size, are: Ministry of Health, Ministry of Public Works, Ministry of Social
Development, and Ministry of Energy.

Data Collection and Data Analysis

This study builds from three sources of information. The most important one is
in-depth interviews, which are coupled with documentation and archival records.
We conducted 51 in-depth interviews; 90% of them were in-person interviews car-
ried out during field work in Chile. To have a homogeneous sample for each case,
we interviewed officials with similar job descriptions, including program managers,
division chiefs, and the head of management- and budget-related departments. One
important distinction is that Chilean line ministries tend to oversee several execut-
ing agencies (known as servicios p
ublicos), and we focus on one of those agencies
per case. The agencies are the Subsecretariat of Health Care Networks at the
Ministry of Health, the Directorate of Waterworks at the Ministry of Public
Works, the Subsecretariat of Social Services at the Ministry of Social
Development, and the Subsecretariat of Energy at the Ministry of Energy. Finally,
additional interviews were conducted with current and former officials from the
Ministry of Finance, the Ministry General Secretariat of the Presidency, the
National Civil Service Directorate, and with members of the academia and civil
society. All interviewees were granted anonymity.
Data analysis follows process-tracing methodology. Some of the main process-
tracing principles used in this study are: providing an in-depth description of
internal processes guided by a predefined theoretical framework; contrasting alter-
native and/or rival explanations of the mechanisms that influenced behavior and
decision making; and discussing, based on available evidence, which alternative is
the most likely explanation.
Before presenting our findings, we address potential weaknesses by discussing
the external and internal validity of this study. The major weakness of studies from
a single country is their low potential for generalization. With that in mind, we use
a comprehensive conceptual framework that was built from the variables men-
tioned in studies that cover many countries where performance budgeting has been
implemented. By using that framework, we try to ensure that the aspects covered
in this study relate to those that have been found relevant in other cases.
596 International Public Management Journal Vol. 23, No. 4, 2020

Figure 1. Chile’s line ministries by the size of their budgets, clustered, year 2016.

The situation is different with regards to internal validity. Analyzing a single

country study and using process tracing leads to higher understanding about the
characteristics of the case. This study uses the process-tracing methodology for
descriptive and analytical purposes. Process tracing is used to provide in-depth
description of internal processes, such as budget elaboration, budget negotiations,
Performance Information in Chilean Ministries 597

and performance monitoring during budget execution, and to guide the discussion
of the causal mechanisms that explain our findings.
Another aspect that enhances the internal validity of our findings is that the tri-
angulation of evidence shows consistent results. For instance, we repeatedly
obtained accounts from different interviewees, each of them interviewed separately
and anonymously, providing the exact same description of specific processes. In as
many cases as possible, we also contrasted the descriptions from interviewees with
evidence from internal documents. One final caveat to this argument is that almost
two-thirds of our interviewees are from line ministries; therefore, the depiction of
the relations between line ministries and other institutions, such as DIPRES, may
be biased.

FINDINGS
There is evidence that performance information is used in the four cases
included in this study. In all cases, performance information has a similar, yet mar-
ginal, role for budget formulation and negotiation. The most significant differences
across cases arise during budget execution, where some line ministries have imple-
mented elaborated and comprehensive performance monitoring systems, while
other cases show more limited use.

Budget Formulation and Approval

The formulation of the budget for any given year officially starts in the month
of July as DIPRES communicates the budgetary ceilings for each line ministry.
However, informal budgetary work within ministries starts months before that. It
is customary practice within ministries to have a draft budget request ready by the
end of June, just before they are informed of their budgetary ceilings. Program
managers start preparing their draft budgets in May by relying on two types of
data: costs and coverage, where the latter is expressed as the number of people
who will benefit from the program. The total amount of resources tends to be a
function of those two variables following the simple equation: Total Program
Budget ¼ Number of Beneficiaries  Cost per Beneficiary.
In all of the cases analyzed, the costing process is decentralized and informal;
i.e., there are no centralized or systemic mechanisms to collect updated cost infor-
mation. Despite that, no interviewee expressed any concerns about costing practi-
ces nor considered that public officials purposefully manipulated their cost
information.
If the total budget increases exclusively due to higher costs, while keeping cover-
age the same, then the increased budget is considered “continuity” instead of
“expansion.” In other words, a continuity budget is one that allows the program to
maintain its current service level. The distinction is important because requests for
budget continuity are usually approved without much scrutiny, while budget
expansions are analyzed in detail. This approach is similar to Wildavsky’s (1964)
598 International Public Management Journal Vol. 23, No. 4, 2020

depiction of budget incrementalism, with the exception that the budget base in
Chile is not just last year’s budget, but instead a calculation of the resources
needed to maintain current service levels.
While performance information is not often used when preparing draft budgets,
program managers requesting expansions must make sure that they have perform-
ance information available to justify their request. The type of performance
information changes across cases. For example, program managers at the Ministry
of Health tend to collect health indicators directly related to their program, cover-
age data, financial execution, satisfaction surveys from program beneficiaries, and
results from health care tests applied to program beneficiaries. The examples are
very similar for the Ministry of Social Development: coverage data, financial
execution, satisfaction surveys from program beneficiaries, and some additional
indicators related to the impact of the program on its beneficiaries. The Ministry
of Public Works is an outlier on the type of performance information they
consider. This ministry almost exclusively executes public infrastructure projects
and, for projects of that nature, there is no impact on beneficiaries until project
construction is finalized. For that reason, monitoring focuses almost exclusively on
the timeliness of the physical and financial execution.
Line ministries have one month to adjust their budget drafts to the budgetary
ceilings established by DIPRES. The adjustments tend to follow a two-step
process. The first step involves general cuts and the second step includes a deeper
review of the programs that are candidates for budgetary expansion. Interviewees
from all cases agreed that the cuts in the first step are politically motivated and
often linked to President Bachelet’s government plan and to political priorities
from the Minister.
The second step of budget formulation within line ministries is the one where
performance information has a larger role. The discussions at this stage of the
process comprise exclusively those programs that requested expansion (or that are
totally new) and that passed through the political filter. In other words, these
discussions concern those programs that were already determined to be directly
aligned with the political priorities of the Minister (or other relevant high-level
authority) and/or to the priorities established in President Bachelet’s government
plan. The main concern of line ministry budget officials at this stage is to ensure
that any request for program expansion can be defended later when they negotiate
with DIPRES. To accomplish that, line ministry’s budget officials may call pro-
gram managers and/or division chiefs and ask them to defend their case for budget
expansion. Even if they are not called personally, they will at least be required
to submit performance information to defend their case. While for each program
they might discuss some differentiated performance indicators, discussions focus
predominantly on the two indicators that line ministries anticipate that DIPRES
will ask for: coverage data and financial execution.
While the budget formulation processes described earlier in principle apply to
all cases, the situation of the Ministry of Energy is somewhat different. At the start
of the Bachelet administration, the Ministry of Energy defined a strategic plan,
Performance Information in Chilean Ministries 599

called the Energy Agenda, which was used to establish an informal four-year
budgetary blueprint with DIPRES. Interviewees at the Ministry of Energy
explained that budget formulation and negotiation run smoothly because of that
blueprint. Also, interviewees at DIPRES said that multi-year blueprints are
desirable and that they have agreed on those with other institutions not included
in this study.
Line ministries submit their budget request to DIPRES about a month after
being informed of their budgetary ceilings. The negotiation process between
DIPRES and line ministries lasts a month and moves progressively from discussing
aggregate financial information to performance information of specific programs.
The process starts with an analysis of the continuity budget—i.e., the resources
needed to maintain current service levels and meet existing commitments for
the upcoming year—and it ends with an analysis of the programs that are new or
that are requesting budgetary expansion. That final analysis considers distinct
types of performance information, but it is strongly skewed towards coverage
data, financial execution, and linkage with political priorities. DIPRES officials
explained that the reason why financial execution is so important is because of the
“opportunity cost” of resources: for every expansion that they approve, they are
rejecting an expansion to someone else. Therefore, DIPRES officials are highly
reluctant to provide additional funding to programs and line ministries that have
been unable to spend all of their resources in the past, as they are expected to be
more likely to again end up with idle resources. Those same DIPRES officials
admitted that the use of other performance information during budget negotiations
is “suboptimal.”
The Chilean Congress has a very limited role in negotiating the executive budget
proposal. This is arguably a consequence of the fact that the legislature lacks the
legal authority to negotiate the budget. For instance, the Chilean constitutions
bars the Congress from increasing expenditures, reallocating resources among
programs, and amending the economic assumptions or the revenue forecasts
(Bl€ondal and Curristine 2006). Additionally, if Congress fails to act on the budget
by November 30, then the original executive proposal becomes law. Because of this
situation, legislators hold more limited hearings and do not systematically call line
ministries to testify on their budget proposals, but instead DIPRES takes the lead
role of explaining and defending the budget request in representation of the
executive branch. Interviewees at line ministries explained that when legislators do
call them to testify, they almost exclusively focus on specific financial restrictions,
such as personnel hiring caps, and reporting requirements which are included
as footnotes on each institution’s budget.

Budget Execution
The main differences in the use of performance information arise during budget
execution. The case of the Ministry of Energy stands out for the strength of its
institutional-level performance monitoring system. The system was created in 2014
based on their strategic plan—the Energy Agenda—and it introduced the position
600 International Public Management Journal Vol. 23, No. 4, 2020

of a Project Management Officer (PMO). The first steps to create the system were
to analyze the projects within the institution, to examine their linkage to the
Energy Agenda, to define specific goals and deadlines, and to assign people respon-
sible for each goal. These processes resulted in a total of 99 goals and 17 people
responsible for those goals.
The monitoring of those goals is carried out through recurrent weekly and
monthly activities (see Table 2). The first step requires program managers to moni-
tor how decentralized offices are managing program execution and to report results
monthly through an information system. There is one official counterpart in each
division who is responsible for making sure that results are reported in a timely
way. There are also short weekly meetings where division chiefs can raise issues to
the PMO. The main meeting happens once a month and includes division chiefs,
the PMO, the Subsecretary, and the Minister. These meetings start with an over-
view of all goals through a “traffic light” system; continue with a detailed analysis
of the 10 most delayed projects, establishing specific commitments to help put
them back on track; followed by a brief presentation about each division; and end
with a 10-minute overall presentation by the PMO. Finally, there is a parallel sys-
tem of monthly meetings between the PMO and decentralized offices focusing on
specific goals that were established for the latter.
The reports used in the monthly meetings are created using the information
from the two systems that were built when the PMO office was instituted: The
Project Management System, which includes program execution and deadlines in a
Gantt chart, and the Comprehensive Command Chart, which has performance

TABLE 2
Institutional-Level Performance Monitoring at the Ministry of Energy

Weekly Monthly
Steering Committee Meetings using “traffic light”
system to identify progress.
Definition of commitments
for next meeting. Establish
action plans to meet
commitments.
Division Chiefs Short weekly meetings to
discuss additional issues.
Meet with PMO, at least
before each monthly
meeting, to review what
will be discussed.
Decentralized Offices Provide information for Monthly review of
project reporting. decentralized goals.
Project Managers Input project information to
the management system.
Performance Information in Chilean Ministries 601

indicators. Interviewees explain that the information discussed in these meetings

often leads to resource reallocation between program components and geographic
locations, and to mobilization of resources to evaluate the causes behind anomalies
that are impacting program execution. In addition, interviewees also expressed that
monitoring meetings serve as motivation to improve and showcase their
performance.
The institutional-level performance monitoring system at the Ministry of Public
Works, and at one of its agencies, the Directorate of Waterworks, is also highly
institutionalized within the organization. The system includes two separate tiers.
The higher tier involves six executing agencies under the Ministry of Public Works,
one of which is the Directorate of Waterworks. There are weekly meetings (every
Monday) that are very short, attended by the head of each executing agency, and
focus on the short- and medium-term horizon of physical and financial execution
(no more than three months ahead). For each meeting, there is a report, prepared
with information from the Integrated Financial Information System (SAFI), which
is an Excel-based platform that has project-level information with comparative
data of the physical and financial execution. Interviewees who attend these meet-
ings consider them to be positive and tend to highlight two aspects: that they gen-
erate competition among executing agencies and that they emphasize explicit
commitments that are followed up on the next meeting.
The second tier of institutional-level performance monitoring is separate for
every executing agency. In the Directorate of Waterworks, the system is led by two
PMOs, each in charge of monitoring diverse types of projects and each allowed to
tailor processes according to their needs. Both PMOs are responsible for getting
updated information, setting a periodic meeting calendar, and preparing perform-
ance reports for those meetings. The main source of information for those reports
is the same SAFI system used for the weekly executive meetings described in the
previous paragraph. The information discussed in each meeting is used to certify
that contractors are meeting their contractual obligations, to determine whether to
grant financial resources requested by decentralized project members, to take man-
agerial and budgetary actions to ensure meeting physical and financial execution
milestones, and to set specific commitments and deadlines that are monitored
subsequently.
The other line ministries analyzed in this study have also implemented perform-
ance monitoring systems. For instance, the Subsecretariat of Health Care
Networks, which belongs to the Ministry of Health, has three performance moni-
toring systems at the institutional level. All of those systems have a formulaic
approach that is used to award financial transfers and/or monetary incentives. The
first system is known as the Managerial Commitments (COMGES) and consists of
a set of strategic goals, each measured by one or more indicators. The second sys-
tem is the Primary Health Care Index (IAAPS), which was established in 2002 and
provides a monetary incentive linked to a set of goals for local service centers. The
goals, known as Sanitary Goals, are set each year by the Subsecretariat of Health
Care Networks and are almost entirely composed of coverage indicators. The third
602 International Public Management Journal Vol. 23, No. 4, 2020

system is known as the Customer Service Bonus. This system was created in 2012
and provides a monetary incentive to personnel from local service centers. The
monetary incentives are awarded based on the results of a customer satisfaction
survey designed by the Subsecretariat of Health Care Networks. Health centers
with a survey score of 65% or higher are eligible for the monetary incentive.
Despite having three institutional-level monitoring systems, interviewees at the
program level suggested that those systems have no impact on their day-to-day
activities. This is a consequence of the fact that these systems were not built around
individual programs, but instead they focus on the multiple health care centers that
the Subsecretariat of Health Care Networks supervises. This study did not cover
the supervisory role of the Subsecretariat of Health Care Networks where these
systems are likely to be relevant.
Similarly, most interviewees at the Ministry of Social Development did not men-
tion the monthly institutional-level performance monitoring system as relevant.
One particularity of the monitoring system at the Ministry of Social Development
that helps explain its lack of success is that it is the only system that focuses on the
performance targets established with DIPRES, instead of on targets defined exclu-
sively within the institution. Additional explanations for the relative success of per-
formance monitoring systems at the Ministry of Energy and the Ministry of Public
Works, as compared to the other two cases, are analyzed in the discussion section
of this study.
Finally, in all cases, institutional-level monitoring is complemented with that at
the program level. The core monitoring activities described by program managers
are similar for all cases. Omitting some minor differences, those activities are:

 Engage daily with decentralized offices to gather information about pro-

gram execution;
 Tabulate the information into an information system that is exclusive for
that program;
 Analyze the information in the system in search of anomalies;
 Request additional information to confirm that the data are correct;
 React to those anomalies through decisions that might involve changes to
service-delivery processes, reallocation of resources within between-program
components and geographic locations, among others.

The main difference in program monitoring across institutions is the informa-

tional system capacity. This capacity affects what type of indicators are readily
available, how much time program managers must spend in data revision, and how
easily they can analyze larger sets of information. Almost all of the interviewees at
the program level expressed the need for better individualized systems. The most
frequent practice is to work with ad-hoc Excel files that are, at least in some cases,
constantly transferred by multiple users through emails. Those who were satisfied
with their current informational system capacity were either from the Ministry of
Energy, where the small size of programs allows for simpler systems, or from the
Performance Information in Chilean Ministries 603

Ministry of Public Works, where most program managers focus on the same
indicators already included in the institutional system.
Besides differences in their informational system capacity, program monitoring
differs in the frequency of internal reporting. Among those interviewed, there
where some who requested reports to decentralized units every two, three, six and/
or 12 months. In all cases, the reports include more detailed performance informa-
tion compared to what they analyze in their everyday work.
Besides using data from institutional- and program-level monitoring, there were
two program managers who reported using performance results from ex-post eval-
uations. The most prominent case is the Decent Night program, from the Ministry
of Social Development, where the findings of the ex-post evaluation motivated a
comprehensive reform of one of the program’s components and the creation of a
system of working meetings between the Ministry of Social Development and three
other institutions whose work is relevant to the program.
While performance monitoring leads to changes within programs, performance
information is not a major factor in deciding budget reallocations across programs.
Interviewees from all cases confirmed that reallocations across programs are based
exclusively on financial execution; i.e., resources are transferred from under-execut-
ing programs to those that could use additional funds.
The internal finance division of each institution has the responsibility of moni-
toring financial execution to promptly identify programs that are lagging. Budget
officials at DIPRES do their own monitoring of financial execution by contacting
line ministries and discussing, at least twice a month, the potential need to reallo-
cate resources. The reallocation of funds among programs can be burdensome due
to administrative controls. For example, interviewees from the Directorate of
Waterworks explained that reallocating funds among investment projects might
take between three and five months, as they need the approval of the Ministry of
Finance and the Comptroller and, depending on the magnitude, it might require
redoing ex-ante evaluations.
Reallocations within programs are simpler and might be based on performance
information. This is at least partly because the Chilean budget is approved on
aggregated accounts, without providing much detail to the legislature on the spe-
cific line items within programs. For instance, one program manager reallocated
idle funds after the price of the main input for his program, which was costed
based on information from two years before, dropped significantly. The program
manager analyzed available data and decided that the additional funds should be
used to expand program coverage.

DISCUSSION
In this section, we address the two main questions that arise from the findings
of this study: (1) Which factors explain the relative success in the use of perform-
ance information found at Energy and Public Works?; (2) Why do differences
across line ministries arise almost exclusively during budget execution?
604 International Public Management Journal Vol. 23, No. 4, 2020

This study shows that the performance monitoring systems developed and insti-
tutionalized at the Ministry of Energy and the Ministry of Public Works are more
robust. Those two cases have some characteristics that differentiate them from the
other two cases included in this study: higher internal leadership support and more
adequate informational system capacity. In addition, the systems that these two
government agencies installed have the common particularity of including a PMO.
Finally, there are some particular characteristics that might facilitate (or hinder)
success. We discuss these factors in the following paragraphs.
Many interviewees from the most successful cases said that the fact that the
internal high-level leadership cares about performance has been critical for the suc-
cess of their current monitoring systems. Those same interviewees suggested that
lower turnover rates of their leadership have been important for getting them
involved with the details of individual projects and with the strategic goals of the
organization. In fact, some of those interviewees, who have been in their organiza-
tions for several years, argued that the current situation is new and that, in previ-
ous years, they did have high turnover and weak leadership involvement. These
findings contrast directly with that from the least successful cases, where interview-
ees linked recent high turnover rates and re-organizations with lack of clarity on
strategic goals and diminished executive involvement in program success and per-
formance measurement.
Another aspect in which the most successful cases have a better situation is
related to informational systems. Two-thirds of those interviewed at the Ministry
of Energy said that they have access to quality and timely information for their
decision-making processes due to appropriate informational systems, while the
other third said that they have access to most, but not all, of the information they
need. While the proportion of positive respondents drops to 42% at the Ministry of
Public Works, it remains much higher than that of the remaining two cases, where
all program managers reported multiple issues with their informational systems,
leading to less availability of information, more time spent in data revision, and
difficulties analyzing larger sets of information.
There are several reasons why the Ministry of Energy and the Ministry of
Public Works have more adequate informational systems. The situation at the for-
mer is the result of the explicit interest of its leadership in instituting a robust per-
formance monitoring structure for their Energy Agenda. In addition, the situation
was facilitated by its small size: program managers at the Ministry of Energy do
not require advanced software to manage their programs because they are small
enough to be monitored through ad-hoc Excel spreadsheets. The case of the
Ministry of Public Works is also explained by an intrinsic advantage: they have
very homogeneous projects (almost exclusively public infrastructure) that use simi-
lar performance metrics and software to track most of their portfolio. For that rea-
son, program managers and those in leadership positions get almost everything
they need with a single informational system.
It is important to state that there have been efforts at the remaining two cases
to implement informational systems. In both cases, there are broad systems that
Performance Information in Chilean Ministries 605

provide general information that is useful for some programs. Also, at the time of
the interviews, one program in each of those institutions had already contracted a
firm to develop an individualized system for them. However, programs at these
two institutions are neither small nor homogeneous, so their needs are harder
to satisfy.
In fact, size was not only an advantage for the Ministry of Energy, but also a
disadvantage for the Ministry of Health. As mentioned in the previous section,
the Subsecretariat of Health Care Networks, which belongs to the Ministry of
Health, has implemented three performance monitoring systems at the institutional
level. The reason why interviewees at the program level do not find those systems
too relevant for their jobs is that those systems were designed to have an aggre-
gated, instead of program-related, focus. Having aggregate systems might have
been a correct decision, given that the Subsecretariat of Health Care Networks is a
very large executing agency and that its leading role is to coordinate many decen-
tralized health care units. However, that approach left many program managers
without adequate informational support. This shows how a larger organization has
a more complex task when designing comprehensive performance monitor-
ing systems.
A final difference between more and less successful cases is the creation of a
PMO. In the most successful cases, the PMOs played a pivotal role in performance
monitoring. For example, PMOs established a meeting structure, generated reports,
presided over meetings, and monitored program and division managers, among
other functions. The role of the PMOs is broad in terms of performance improve-
ment: the information they collect and the meetings over which they preside serve
to decide on budgetary (such as the reallocation of funds within and between pro-
grams) and managerial matters. In the least successful cases, there seems to be no
position equivalent to PMOs.
In sum, the evidence presented here points towards three main success factors.
First, having executive leaders who actively care about performance monitoring,
where this factor is highly dependent on having lower managerial turnover rates.
Second, having adequate informational systems, which can be facilitated or
obstructed by some intrinsic characteristics of the organizations. And third, taking
the right approach in designing monitoring systems, such as establishing clear
responsibilities (such as those of a PMO office) and establishing a structured and
periodic meeting system. These success factors are complementary, rather than
competing, explanations, and while all of them might have had an impact, it is dif-
ficult to determine their relative importance.
We now move to the second question of this section: why do differences among
line ministries arise almost exclusively during budget execution? One fact that
stands out from this study is that budget formulation, negotiation, and approval
are almost identical for all cases. The most likely explanation is based on one of
the characteristics of Chile’s budget formulation, negotiation, and approval: they
are heavily centralized and dominated by DIPRES, leaving almost no room for
line ministries to maneuver.
606 International Public Management Journal Vol. 23, No. 4, 2020

One piece of evidence that supports this argument is the fact that line ministries
have adapted their internal discussions to base them on the same types of informa-
tion that DIPRES asks during negotiations: coverage data, financial execution,
and political priorities. This was confirmed by all program managers, by other
interviewees at line ministries who have been involved in budget negotiations with
DIPRES, and by DIPRES officials themselves. The fact that all cases focus on
that same information, and that the information matches what is requested by
DIPRES, suggests that they are all being influenced by the same phenomenon.
DIPRES has taken a different approach to how it manages its relationship with
line ministries during budget execution. The control tools used by DIPRES for
ensuring administrative compliance during budget execution have a very delimi-
tated scope. On financial terms, DIPRES only oversees reallocations across aggre-
gated accounts while providing flexibility for transfers within programs. This
contrast is illustrated by two examples provided earlier in this study: the long and
tedious process that line ministries go through when reallocating resources across
projects and the ease with which a program manager could reallocate excess
resources from cost efficiencies. On performance terms, DIPRES focuses exclu-
sively on the indicators from their centralized systems. An overwhelming majority
of interviewees at line ministries explained that indicators from DIPRES systems
have a very limited impact on everyday program management. These looser finan-
cial and performance controls have allowed line ministries to develop more useful
and differentiated monitoring processes.
The fact that DIPRES has a less prominent role during budget execution might
also reduce the incentive for line ministries to obscure negative information. For
instance, while showing evidence of low performance during budget formulation
might result in reduced funding from DIPRES and in no gains for the ministry,
using honest performance information during budget execution involves no exter-
nal penalty and may potentially lead to more successful interventions.
However, the role of DIPRES is not the sole reason why line ministries have
had more liberty to establish customized processes during budget execution. One
major difference between budget execution and budget formulation is how time-
constrained they are. On the one side, budget formulation and approval are highly
limited by time pressures. For example, line ministries only have 30 days to adapt
their budget requests to their budgetary ceilings. On the other side, budget execu-
tion lasts at least 11 months every year—interviewees explained that the month of
January is almost null for budget execution—providing ample time to establish
monitoring systems based on frequent reports and meetings.
The claim that tight timeframes are an obstacle to using performance informa-
tion is not new. In fact, it aligns closely to the well-known concept of budgetary
incrementalism. One of the major arguments of incrementalism is that public offi-
cials do not have the time (or capacity) to analyze and compare competing alterna-
tives (Lindblom 1959; Wildavsky 1964). The decision-making processes within
Chile’s government agencies during budget formulation and negotiation fit the idea
of incrementalism, as decisions are based on a narrow amount of information
Performance Information in Chilean Ministries 607

(mostly financial execution and political priorities), while additional analysis is

only used for marginal changes (programs seeking budgetary expansions).
However, it should be noted that incrementalism is somewhat “smarter” in Chile,
as the budget base is determined by the financial resources needed to keep current
service levels (continuity budgets), and not simply by the financial resources allo-
cated for the previous year.
A final alternative explanation for the differences between budget formulation
and execution is that performance information is more relevant during budget exe-
cution because program managers are naturally more performance-oriented than
budget officials. While this might be true, there are at least two issues with this
argument. First, reports from previous authors suggest that DIPRES budget ana-
lysts had a more performance-oriented focus during the Frei and Lagos adminis-
trations (Bl€ ondal and Curristine 2006; Guzm
an 2005; Rojas et al. 2005; Zaltsman
2009). Second, there might be an endogenous relation as program managers might
be more performance-oriented because they have more time and flexibility to use
performance information.
In sum, we discuss three potential explanations for the fact that line ministries
have developed processes to use performance information for budget execution
and not for budget formulation. These three explanations are not mutually exclu-
sive; i.e., the presence of one does not preclude the others from being true. The
two most convincing explanations are that budget execution has less strict time-
frames and less bureaucratization from DIPRES. The former provides line minis-
tries more time to analyze and decide based on performance information, while the
latter gives more flexibility to design appropriate systems and reduces the incentive
to obscure negative information.
There is one final clarification about the variables that has not been mentioned
in this discussion section. This study has low degrees of freedom for inferential
leverage, which is common in qualitative studies with limited number of cases and
multiple explanatory variables (King, Keohane, and Verba 1994). The analysis of
degrees of freedom in qualitative studies is not as straightforward as in quantitative
studies where one simply compares the number of cases (or data-set observations)
with the number of variables. Instead, the inclusion of key variables in qualitative
studies may increase inferential leverage, despite decreasing degrees of freedom
(Collier, Brady, and Seawright 2010). The increased inferential leverage is the result
of a better understanding of the causal mechanisms that explain the findings of a
case study, where the better understanding is achieved through the additional
“causal-process” information provided by key variables. The fact that some varia-
bles are not explanatory for our research questions does not mean that they are
not relevant. A good example is the capacity of line ministry staff to interpret and
use performance information. This variable was not found to have explanatory
power because its value was homogenous in all of our cases. The correct interpret-
ation for this and other variables is not to mark them as irrelevant, as one might
do in a quantitative analysis, but to label them as having been held constant for
this study.
608 International Public Management Journal Vol. 23, No. 4, 2020

CONCLUSIONS
The objective of this study is to enhance understanding of performance budget-
ing at the line ministry level. From that premise, we set three research objectives:
(1) understand the types of decisions, particularly during budget execution, for
which line ministry officials use performance information; (2) compare line minis-
tries within the same government to discover potential differences in the way that
performance information is used; and (3) analyze the factors that explain such
differences.
On the first objective, the evidence shows that all of the line ministries included
in this study make extensive use of performance information during budget execu-
tion, both for budgetary and for managerial purposes. While varying in their levels
of success, all line ministries implemented a performance monitoring system that
included a defined set of indicators that were periodically reviewed. In addition, all
program managers established their own procedures for reporting and analyzing
performance information.
In contrast, performance information is only marginally used for resource allo-
cation. During budget formulation and negotiation, performance information is
rarely reviewed for programs that are not requesting budgetary expansions. The
performance review of the remaining cases is very limited and focuses mostly on
program coverage and financial execution.
On the second and third objectives, the study contrasts more and less successful
cases. While all institutions have created processes to review performance, those
processes are more impactful at the Ministry of Energy and the Ministry of Public
Works. Among the characteristics that explain their success: lower managerial
turnover rates, which are correlated with higher managerial involvement in organ-
izational performance; more adequate informational systems that facilitate the col-
lection, use, and storage of performance information; better design of the structure
of their performance monitoring systems; and intrinsic characteristics of organiza-
tions, such as having homogenous programs and/or being smaller in size.

ACKNOWLEDGMENTS
The author is grateful to the Centro de Pol
ıticas P

ublicas from the Universidad
Cat
olica (Chile), and particularly to its Director Ignacio Irarr
azaval, for their
valuable support during field data collection.

REFERENCES
Barzelay, M., and F. Thompson. 2006. “Responsibility Budgeting at the Air Force Materiel
Command.” Public Administration Review 66(1):127–38.
Berry, F. S., R. S. Brower, and G. Flowers. 2000. “Implementing Performance
Accountability in Florida: What Changed, What Mattered, and What Resulted?”
Public Productivity & Management Review 23(3):338–58.
Performance Information in Chilean Ministries 609

ondal, J. R., and T. Curristine. 2006. “Budgeting in Chile.” OECD Journal on Budgeting
Bl€
4(2):7–45.
Cain, C., E. Choudhury, and J. C. Clingermayer. 2004. “Turnover, Trust, and Transfers: An
Examination of Local Government Budget Execution.” International Journal of Public
Administration 27(8–9):557–76.
Cheung, A. B.-L. 2006. “Budgetary Reforms in Two City States: Impact on the Central
Budget Agency in Hong Kong and Singapore.” International Review of Administrative
Sciences 72(3):341–61.
Collier, D., H. E. Brady, and J. Seawright. 2010. “Sources of Leverage in Causal Inference:
Towards an Alternative View of Methodlogy.” Pp. 161–99 in Rethinking Social Inquiry:
Diverse Tools, Shared Standards, 2nd ed., edited by H. E. Brady and D. Collier.
Lanham, MD: Rowman and Littlefield.
Dean, P. N. 1986. “Assessing the Performance Budgeting Experiment in Four Developing
Countries.” Financial Accountability and Management 2(1):1–24.
Dougherty, M. J., K. A. Klase, and S. G. Song. 2003. “Managerial Necessity and the Art of
Creating Surpluses: The Budget-Execution Process in West Virginia Cities.” Public
Administration Review 63(4):484–97.
Downs, A. 1966. Inside Bureaucracy. Santa Monica, CA: Waveland Press, Inc.
Garc
ıa L
opez, R., and M. Garc
ıa Moreno. 2010. La Gesti
on Para Resultados En El
Desarrollo. Washington, DC: Inter-American Development Bank.
Gilmour, J. B., and D. E. Lewis. 2005. “Assessing Performance Budgeting at OMB: The
Influence of Politics, Performance, and Program Size.” Journal of Public Administration
Research and Theory 16(2):169–86.
Gisselquist, R. M. 2014. “Paired Comparison and Theory Development: Considerations for
Case Selection.” PS: Political Science & Politics 47(02):477–84.
Guzm
an, M. 2005. Sistema De Control De Gesti
on y Presupuestos Por Resultados: La
Experiencia Chilena. Santiago, Chile: Ministerio de Hacienda.
Hackbart, Merl, and Jim Ramsey. 1999. “Managing Public Resources: Budget Execution.”
Journal of Public Budgeting, Accounting & Financial Management 11(2):258–75.
Ho, A. T.-K. 2011. “PBB in American Local Governments: It’s More than a Management
Tool.” Public Administration Review 71(3):391–401.
Ho, A. T.-K. 2018. “From Performance Budgeting to Performance Budget Management:
Theory and Practice.” Public Administration Review doi: 10.1111/puar.12915.
Joyce, P. G. 2003. Linking Performance and Budgeting: Opportunities in the Federal Budget
Process. Arlington, VA: IBM Center for the Business of Government.
Joyce, P. G. 2011. “The Obama Administration and PBB: Building on the Legacy of Federal
Performance-Informed Budgeting?” Public Administration Review 71(3):356–67.
King, G., R. O. Keohane, and S. Verba. 1994. Designing Social Inquiry: Scientific Inference
in Qualitative Research. Princeton, PA: Princeton University Press.
Lindblom, C. 1959. “The Science of Muddling through.” Public Administration Review
19(2):79–88.
Mak
on, M., and M. Varea. 2015. “Results-Based Budgeting.” Pp. 75–109 in Building
Effective Governments: achievements and Challenges for Results-Based Public
Administration in Latin America and the Caribbean, edited by J. Kaufmann, M. Sangin
es,
and M. Garc
ıa Moreno. Washington, DC: Inter-American Development Bank.
Mart
ınez Guzm
an, J. P. n.d. “Is Performance Information Used Under Performance Budgeting?
A Systematic Review” International Journal of Public Sector Performance Management.
610 International Public Management Journal Vol. 23, No. 4, 2020

McCaffery, J., and J. E. Mutty. 1999. “The Hidden Process of Budgeting: Execution.”
Journal of Public Budgeting, Accounting & Financial Management 11(2):233–57.
Melkers, J., and K. Willoughby. 2005. “Models of Performance-Measurement Use in Local
Governments: Understanding Budgeting, Communication, and Lasting Effects.” Public
Administration Review 65(2):180–90.
Moynihan, D. P. 2006. “What Do We Talk about When We Talk about Performance?
Dialogue Theory and Performance Budgeting.” Journal of Public Administration
Research and Theory 16(2):151–68.
OECD. 2013. “Performance Budgeting.” Pp. 94–95 in OECD, Government at a Glance 2013.
Paris: OECD Publishing.
Pitsvada, B. T. 1983. “Flexibility in Federal Budget Execution.” Public Budgeting Finance
3(2):83–101.
Rhee, D.-Y. 2014. “The Impact of Performance Information on Congressional
Appropriations.” Public Performance & Management Review 38(1):100–24.
Rojas, F., K. Mackay, Y. Matsuda, G. Shepherd, A. del Villar, A. Zaltsman, and P.
Krause. 2005. Chile: An
alisis Del Programa De Evaluaci
on Del Gasto P
ublico.
Washington, DC: The World Bank.
Ryu, J. E. 2013. “Performance-Based Budgeting (PBB), Allocative Efficiency, and Budget
Changes: The Case of the U.S. Department of Commerce.” Public Finance and
Management 13(4):335–60.
Schick, A. 1964. “Control Patterns in State Budget Execution.” Public Administration
Review 24(2):97–106.
Shah, A., and C. Shen. 2007. “A Primer on Performance Budgeting.” Pp. 137–76 in
Institutions, Budgeting and Budgetary, edited by A. Shah. Washington, DC: The World
Bank.
Thompson, F., and W. Zumeta. 1981. “Control and Controls: A Reexamination of Control
Patterns in Budget Execution.” Policy Sciences 13(1):25–50.
Wang, X. 1999. “Conditions to Implement Outcome-Oriented Performance Budgeting:
Some Empirical Evidence.” Journal of Public Budgeting, Accounting & Financial
Management 11(4):533–52.
Wildavsky, A. 1964. The Politics of the Budgetary Process. Boston and Toronto: Little,
Brown, and Company.
Zaltsman, A. 2009. “The Effects of Performance Information on Public Resource
Allocations: A Study of Chile's Performance-Based Budgeting System.” International
Public Management Journal 12(4):450–83.

ABOUT THE AUTHOR

Juan Pablo Mart
ınez Guzm
an ([email protected]) is an assistant research profes-
sor at the School of Public Policy at the University of Maryland–College Park. His
research examines the factors that influence decision-making processes and the
impact of policy decisions within the realm of public budgeting and finance. Most
of his work focuses on public sector reforms in governments from Latin America
and the Caribbean.