Draft as of September 15, 2018

GUIDELINES FOR PUBLIC EXPENDITURE REVIEWS IN THE HUMAN DEVELOPMENT SECTORS

OFFICE OF THE CHIEF ECONOMIST, HDNVP

JANUARY 2004

AUTHORS

DINA N. ABU-GHAIDA

SUE E. BERRYMAN

DOV CHERNICHOVSKY

MARGARET ELLEN GROSH

MATTIAS K. A. LUNDBERG

TABLE OF CONTENTS

Preface

I.Issues in Public Expenditure Reviews Common to the HD Sectors

II.PER Guidelines for Education

III.PER Guidelines for Health

IV.PER Guidelines for Social Protection

PREFACE

These guidelines for conducting public expenditure reviews (PERs) in the human development sectors (education, health, and social protection) have been written because the Bank’s effectiveness in the HD sectors depends partly on a grasp of how these sectors are funded.

  • PERs are critical to fulfilling the Bank’s mission of poverty reduction.
  • As a provider of financial resources, the Bank is contributing to a government’s budget and thus needs to understand how that government is using its money.
  • Government’s budget management affects the impact of Bank lending.

As a means for the Bank to exercise due diligence, PERs have become a Bank requirement, but PERs affect the Bank’s investments in HD in several other ways.

Effectiveness of Bank lending and non-lending services in HD. Targeting what systems analysts call “root causes” of problems is the most efficient way to get them resolved. Countries’ expenditure patterns for HD services can put in motion incentives that have large and perverse effects. Every analytically strong PER for an HD sector looks for distortions in allocations in the sector—e.g., between levels of education or among income quintiles—and for the incentives that expenditure patterns and budget processes set up. For example, Croatia funds higher education students in ways that do not create incentives for them to complete their studies expeditiously. As a result, only a third of those that enroll ultimately complete their studies. Of those that complete, those in four-year degree programs finish in an average of seven years; those in two-year programs, in an average of five years.

PERs as analytic underpinning for other Bank activities. Increasingly the Bank’s focus on poverty reduction is taking the form of poverty reduction strategy credits (PRSCs) and medium term expenditure frameworks (MTEFs). PERs provide analytic underpinning for both.

Our HD clients need the knowledge yielded by HD PERs in order to compete effectively in the contestable budget-setting exercise that should underlie a MTEF and the priority-setting exercise of a PRSC.

PRSCs are expected to constitute an increasing share of Bank lending to a country. All HD sectors are key to poverty reduction. If we are left out of PERs, we are either left out of PRSCs (or SACs) or come to the table at a significant analytic disadvantage. We are much less able to use PRSC or SAC opportunities to reform the HD sectors because we are less able to identify priority triggers and conditionalities.

Very limited Bank investment in HD sector work. The Bank has been investing relatively little in the analyses of the HD sectors that are key to the intelligent design of lending and non-lending services. PERs are an avenue for building knowledge about the financing, efficiency, equity, and quality of HD services. Obviously, PERs and sector work are not the same. Sector work usually includes topics not addressed by a PER, and sectoral issues, such as educational equity, usually involve more than the implications of the public management of the budget. However, a great deal can be learned about an HD sector through the lens of a PER.

This document has four sections. Section I identifies issues common to the three HD sectors. Sections II-IV present Guidelines for doing a public expenditure review in education, health, and social protection, respectively.

SECTION I. GUIDELINE FOR PUBLIC EXPENDITURE REVIEWS IN THE HD SECTORS

A PER is sector work through a fiscal lens.

I.1.SCOPE OF THE PER

Decide the issues or topics to be addressed. The sector-specific Guidelines of topics in sections II-IV are useful for making this decision.

Be clear about why you are including some topics and not others. For example:

  • This is a participatory PER, and the client wants help on some, but not other, topics.
  • The budget and/or time frame for the work are very limited, forcing a more restricted treatment than might have been desirable.
  • The data needed to analyze a particular topic are missing and cannot be collected in the time frame for the PER.
  • A comprehensive PER for HD sector in question has been completed recently, and this is simply an update. In this case, the earlier work can be referenced.

 Check that the scope fits the purposes for the overall PER exercise. (The task team leader for the whole PER should clarify the objectives for the team.) Examples of different objectives:

  • Contribute to a longer-term strategy for public expenditure analysis and support.
  • Meet the Bank’s due diligence obligations.
  • Initiate/move the policy dialogue in particular sectors and on particular topics forward.
  • Underpin strategy development (e.g., next CAS, PRSP, MTEF, the client’s sector strategy for meeting millennium development goals).
  • Underpin lending operations.
  • Develop the client’s public expenditure management capacity and knowledge.

I.2.ORGANIZING THE KNOWLEDGE BASE

Identify the quantitative and qualitative in-country data that you need, given the issues that you are addressing. The HD-specific Guidelines will help you here. (Work with the Bank’s Resident Mission to identify in-country consultants that can assemble these data prior to your mission.)

Five types of surveys are particularly useful for “following the money”. Find out if any of the following surveys exist for your country. How recent are they? What is their analytic status—for example, if you want to do a benefits incidence analysis and there is an LSMS or household budget survey, have the poverty profiles been created?

  • Living standards measurement surveys (LSMS) and household budget surveys. These let you estimate the incidence of poverty, private expenditures for health and education, and through benefits incidence analysis, the allocation of public expenditures in all three HD sectors by poverty quintiles.
  • Public expenditure tracking surveys (PETS). These are designed to provide information about the flow of money among different tiers of government and frontline service facilities. They let you quantify the problem of “asymmetric information” in public spending, as manifested in the leakage of public funds as they flow through levels of government to their ultimate intended beneficiary, such as poor families, schools, or clinics.
  • Quantitative service delivery surveys (QSDS). These take the frontline service provider such as schools as the main unit of analysis. They assess the service delivery system (inputs, outputs), provider behavior, accountability arrangements/performance measurements in service delivery, and provide baselines for examining the impact of policy and institutional reforms.
  • Labor market surveys. These can be used to assess issues such as employment and wage returns to different amounts of education and characteristics of the unemployed.
  • Anti-corruption surveys. These surveys can measure bribes and other “gray” or “illegal” payments that may or may not be picked up by an LSMS—for example, payments for drugs that are supposed to be provided free or bribes to a university faculty to gain admission.

Work closely with the member of your PER team that is evaluating the government’s budget formation and execution processes. Usually a PER discusses these processes cross-sectorally because they are common across the sectors. However, you need to track what your colleague is finding to determine if these processes are affecting your sector in unique and important ways.

Spend some time checking with colleagues and the Bank’s websites to locate studies, methods, and data that might help you. For example, look at the World Development Indicators for comparators; the Bank’s external website under Research and Publications; or, if an issue such as decentralization is a big issue for your country, with appropriate thematic groups for guidance (e.g., PREM’s thematic group on decentralization.) DEC is a good source of information on state-of-the-art methods.

Time and budget are adversaries of what we should all do: check for relevant studies and data from outside of the Bank. Try the OECD website; go to JOLIS through the Bank’s web page, selecting databases of interest such as EconLit; have the Bank’s resident mission collect within-country studies from universities, NGOs, donors.

When key data are missing:

  • If time and budget permit, organize new data collection (possibly in collaboration with other donors). For example, you may need data best provided by a quantitative service delivery survey (QSDS) or a public expenditure tracking survey (PETS). Go on the Bank’s website to get descriptions of these two instruments, examples of their use, and Bank staff expert in their conduct. (World Bank > Research > Program > Public Services > Tools: Frontline Service Provider Surveys.) Survey Tools for Assessing Performance in Service Delivery by Jan Dehn, Ritva Reinikka, and Jakob Svensson is a useful overview paper on the web in a PDF file.
  • If time and budget do not permit, flag studies that are high priority in future.

I.3.SELECTION OF COMPARATORS AND BENCHMARKS

One of the most important contributions of a PER is to estimate “how much is enough”. When are they “too much,” “too little,” or “about right”? There are three sources of benchmarks to guide these judgments:

  • Empirical literature that relates policies in the sector to the outcomes desired, such as enrollment rates, completion rates, and learning achievements.
  • Comparisons with other countries in some way “like” the country in question.
  • Country-specific standards.

Empirical literature. Empirical studies, especially those that relate inputs to outcomes are useful guides. For example, the Program for International Student Assessment, conducted by the OECD, found that a student/teacher ratio of 25:1 was optimal in terms of learning outcomes for 15 year olds. This result implies that most countries have inefficient student/teacher ratios—too low (e.g., most OECD and ECA countries) or too high (e.g., most countries in Africa).

However, such studies have two problems. Often there is no critical mass of studies that can be used to check for consistent findings about relationships between particular inputs and outcomes of interest. Second, there is the “unmeasured variable problem”. Country-specific factors that may or may not be represented in completed studies can alter the relationships between inputs and outcomes.

Comparisons with other countries. Comparisons with other countries have other problems. For example, in the Bank it is very common to compare the percent of GDP or total public expenditures devoted to the social sector to that devoted by countries “comparable” in some way—regional neighbors or per capita GDP, for example. There are two problems. One is that regional neighbors may share the same difficulties—e.g., inefficiencies—as the country in question. For example, countries of the former Soviet Union had the same inefficient input norms, and several continue to use these norms. The second problem is that countries differ in several factors that significantly affect total expenditures in the sector, such as variations in:

  • number of service beneficiaries—e.g., the number of families that need social assistance or the number and enrollment rates of school age children whose education has to be financed;
  • prices for key inputs, such as doctors or teachers;
  • residential patterns that determine opportunities for economies of scale—all else equal, it costs more to provide health care or education in countries with large numbers of small and isolated settlements (e.g., Kazakhstan);
  • policies on public versus private financing.

Country-specific standards. The final and ultimately preferable option for judging how much is enough is developing benchmarks for the country itself, triangulating among the fiscal implications of the country’s goals for the sector; its demographic patterns; analyses of its expenditure patterns; evidence on how efficiently inputs are used, given the country’s settlement patterns and other factors; results of international studies of input-outcome relationships; and data on reasonable comparator countries.

I.4.QUALITY OF THE ANALYSES

Statistical methodologies: if in doubt about the state of the methodological art, check with your Network or DEC.

Root causes: PERs in the HD sector tend to be particularly weak in identifying the incentives that are producing observed expenditure patterns. For example, one PER stated that a particular province needed more money for education, noting that the province received only 60 percent of the national average in Federal expenditures per student. However, the author failed to ask why the province was receiving less than the national average. It therefore failed to evaluate whether the funding shortfall for the sector could have been remedied by re-negotiating the province’s share of Federal transfers. Efficient recommendations are ones that focus on factors that have big and ramifying effects on the sector.

I.5.RATIONALE FOR AND AGAINST PUBLIC INTERVENTION

The fundamental arguments for public intervention are market failures and social justice.

Market failures include externalities, failures in capital markets, and information asymmetries. Take the education sector as an example. Education is a private good with externalities for the larger society. If education were simply a private good without externalities, the state would have no role. However, in addition to positive benefits captured by the individual, such as higher wages, education also has positive externalities for the collectivity, such as better health practices that limit the spread of infectious diseases, higher rates of voter participation, contributions to economic growth that are not entirely captured by the individual in the form of higher wages, and, depending on occupational structures and labor markets, less inequality in incomes and therefore a lower probability of social unrest.

Because education has externalities, governments regulate its consumption through laws defining compulsory education, fines of parents that do not send their children to school, or truant officers to find absent children. Neither children nor parents are trusted to invest sufficiently in education. Children are too young to make these choices on their own behalf, and, relative to their parents, are powerless to enforce choices that are to their, but not to their parents’, benefit. Parents are assumed to under-invest on the part of the child for any of several reasons.

Providers and students also have information unavailable to the other, and providers have information unavailable to taxpayers, stakeholders, and policymakers. As a result, providers may get worse students and students may end up in worse institutions than each wants. Taxpayers may get worse outcomes from the education system than they want or pay more for what they get than they should. The information role of the state derives from these realities: examination systems that yield objective and meaningful evaluations of students’ competencies and information on the outcome and efficiency performance of institutions and the sector.

Equity problems emerge, not from market failures, but from the fact that markets are not set up to equalize initial endowments. Markets are a mechanism for the exchange of goods and services between producers and purchasers that can pay (currency or barter) or borrow. However, individuals’ variations in their initial endowments affect their abilities to pay or to borrow. This fact, the fact that the state has a social justice or redistributive role to assure that individuals have equitable opportunities, and the fact that markets do not recognize the benefits to individuals that are not party to an economic transaction (externalities) are all grounds for the state to intervene.

The public role can be provision,financing,regulation, and information. Although government almost always has a regulatory and information role, it does not have to provide the service to assure equitable access and quality, and it may finance selectively. This figure shows how countries can vary on these two dimensions.

Degree of Public Financing

100%

Voucher schools, regulated private schools (Chile basic, Netherlands, Belgium) / Public schools with little or no cost recovery (many)
Unsubsidized private sector (Philippines tertiary level) / Public facilities with cost recovery (US, Korea, Chile universities)

0%100%

Degree of Public Provision

Source: Elizabeth King

The fundamental arguments against public intervention are government’s failures to assure equitable access and quality, efficient and effective services, and sustainable financing. Policymakers, elected officials, and service providers are more responsive to the demands of wealthier and elite consumers (the “capture” phenomenon), resulting in resources being focused on these groups. For example, public resources can be concentrated on hospitals at the expense of primary health care facilities in response to the demands of elite consumers.

Elected officials can use resources to secure their power base in future elections—for example, by lavishing capital investments on regions tipping politically toward the opposition. Since doctors and teachers, but not textbooks, vote, they can raise their salaries at the expense of providing inputs such as textbooks that complement the sector’s human resources. For example, in Mexico funding for a program, PRONASOL, was spent disproportionately in areas controlled by the party in power, PRI.