Targets and Indicators for MDGs and PRSPs: What Countries Have Chosen to Monitor

The Development Data Group[1]

Development Economics Vice-Presidency

The World Bank

Ever since the adoption of the Millennium Declaration by the UN General Assembly in September 2000 and the subsequent presentation of the Millennium Development Goals (MDGs) in the Secretary General’s “Roadmap Report” in September 2001[2], there has been debate over how well the MDGs capture the aspirations of developing countries. Because the MDGs were set by a political process conducted through a series of global conferences and summits, there has been a suspicion that some of the goals and the specific, quantified targets and the indicators used to monitor progress toward the MDGs were not, in fact, “owned” by the citizens, development agencies, and local policy makers. It has also been suggested that the MDGs have added to the reporting burden on over-taxed statistical systems in developing countries.

In this paper we compare the targets and indicators of the MDGs with the measurable targets and indicators set by twenty seven countries whose Poverty Reduction Strategy Papers (PRSPs) have been accepted by the Boards of the World Bank and the IMF by April 2003 and published on the PRSP website.[3] The process of arriving at a poverty reduction strategy is anchored by four key elements, which define the contents of a Poverty Reduction Strategy Paper (PRSPs) and set a framework for assessing its effectiveness. A poverty reduction strategy should provide:

  1. A comprehensive understanding of poverty and its determinants;
  2. A set of public actions that have the highest poverty impact;
  3. A transparent process for setting budget and policy priorities;
  4. Outcome indicators that are set and monitored using a participatory process.[4]

If these guidelines are followed, PRSPs should reflect a consensus of views on the development program for a country.

There is no required set of indicators or goals that must be included in a PRSP. They are determined within each country. A participatory process for selecting goals and a reporting process that allows all stakeholders to monitor public actions and assesses progress toward the goals is an integral part of every PRSP. For such a process to be credible and to contribute to a deeper understanding of the poverty reduction process, the outcomes must be measurable using reliable and publicly available indicators.

The Millennium Development Goals are based on a similar conceptual framework. Because the goals incorporated in the Millennium Declaration arose from a series of high level conferences and global summits held over the 1990s, they can be said to represent a diagnosis of the causes of poverty and the agreed priorities for alleviating poverty. The MDGs provide a framework for quantifying many of the desired outcomes and building agreements about the actions for achieving them. The specific targets and indicators associated with the MDGs have been reviewed by a succession of technical working groups made up of experts from international organizations and development agencies. The 8 goals, 18 targets and more than 50 indicators (See annex A) therefore represent a confluence of views among policy makers and development experts on an international development agenda. The conviction underlying MDGs, like the PRSPs, is that the goals should be publicly agreed, should target specific, quantified outcomes, and that progress should be monitorable using reliable and readily available indicators.

Reviewing the PRSPs

By the end of April 2003, 27 completed PRSPs were published and available on the World Bank’s website.[5] Each PRSP was reviewed by a staff member in the World Bank’s Development Data Group. Their instructions were to take note of all articulated goals and specific targets mentioned in the PRSP. Each major goal and its supporting targets and indicators were classified under one of the MDG goals (for example, “Eradicate poverty and hunger”), even if the precise targets and indicators differed from those of the MDGs. In cases where the PRSP goals were not objectively related to an MDG goal, new headings were created (for example. “Develop culture and improve the spiritual life of people”, “Rural development”, “Macroeconomic objectives”). For the 16 earliest PRSP countries, the information for each country was recorded in a separate table. The table also includes information on monitoring plans (monitoring agency, data source, frequency); relevant metadata pertaining to the availability and quality of indicators from General Data Dissemination System[6] (if available); and comments or references made by the Joint Staff Assessment on the country’s monitoring and evaluation plans. During the review process, discussions were held among reviewers to standardize terminology and provide some uniformity of treatment.

To make the information from the 27 country information more readily accessible, a summary table was prepared showing the coverage of the MDG and non-MDG targets and indicators of 27 PRSP countries. The goals, targets, and indicators from the 27 PRSPs have been combined under six, broad categories corresponding to principal objectives of the MDGs:

1) Poverty and Hunger
2) Education
3) Gender
4) Health
5) Environment
6) Global Partnership

Coverage of the PRSPS

Using the combined table, it is possible to look across all 27 countries for any indicator to see whether it was included in a PRSP. However, certain targets and indicators recorded in the country tables have not been included in the combined table. Because the MDGs do not include resource allocation indicators (for example, increasing expenditures on primary education ), these have been omitted. Likewise targets having to do with the approval of a law are not included.. The combined table also does not include indicators for which a recognized measurement method is lacking -- for example “increase the awareness of the need for adequate nutrition for pregnant women.” Annex B contains a copy of the combined table. General rules to select indicators are difficult to determine, and we are still discussing such criteria.[7]


Table 1 summarizes the coverage of major categories of MDG indicators in the 27 reviewed PRSPs. A PRSPs is included in a category if it includes at least one related monitorable indicator for the category. The indicators may not be same as those specified in the MDGs. For example, Vietnam included 4 indicators related to primary education enrollment or attendance in its PRSP monitoring plan, one of which happened to be the net primary enrollment rate. Yemen included only one indicator, the gross primary enrollment ratio. For the purposes of Table 1 each country was included once in the calculation of the coverage rate for primary education.

The review of the PRSPs shows that they encompass most of the MDGs, although the targets and the actual data required for monitoring are often different. For example, all countries mention goals for primary education in their PRSPs, but only one country includes all three MDG indicators for education (net primary enrollment, pupils starting grade 1 who reach grade 5, and youth literacy), 15 have none of them, and only 7 include an indicator to measure primary school completion. But all have at least one indicator for school enrolment/attendance, and 16 have at least one indicator for literacy.

For immunization indicators, 20 countries have some indicators for immunization, but only two countries have plans to monitor the MDG indicator for immunization, the proportion of one-year-old children immunized against measles. Two more countries have plans to monitor a similar indicator. Other related monitoring indicators include immunization programs for DPT, BCG, polio, and yellow fever.

As shown in Table 1, above, the alignment with the MDGs is closest with the specific targets for school enrollment/attendance (100%), maternal health (100%), child mortality (96%), poverty headcount (93%), and access to water (89%). Looking across all indicators specified for each MDG goal, the alignment is closest for health (76%) and with education (69%), but weaker for gender (36%) and environment (37%).[8]

Goals that are often cited by PRSP countries but are not part of the MDGs include: macroeconomic management, rural and/or agricultural development, infrastructure, governance, private sector reform, and social protection.

Comparing targets

The next step in this study was to compare target rates of progress between the MDGs and goals set by PRSPs. For the purpose of comparability, this exercise could only be done in cases where a country has identified a quantified target for an indicator that is the same as or very similar to the MDG indicator. Review of 27 PRSP indicators revealed there are many other, closely related indicators in the PRSPs, which are intended to monitor outcomes related to the same overarching goals, but we have not included them for the target comparison, because the relationship of the PRSP target to the MDG target is not clear (for example, gross enrollment rates, for which many levels are consistent with full primary enrollment) or because of the even more tenuous relationship between "proxy" indicators and targeted outcomes. We have also left out MDG indicators that are so infrequently mentioned that a comparison did not seem useful, for example, $1 a day poverty or measles immunization (which also has no agreed MDG target). There are still many ambiguous cases. For example, we found 11 slightly different indicators identified in PRSPs that are similar to “access to improved water source” including “access to safe drinking water”, “access to potable water”, and “access to clean water.” We have decided to include these indicators, assuming that they measure conditions that are similar enough to make the comparison valid. But this decision may have to be revised once we collect information about the exact definition of the PRSP indicators. We also included infant mortality rates and assessed them using the same rate of change as the under-5 mortality rates, although there is technically no MDG target for infant mortality. Also, we have included primary completion rates along with net enrollment rates. Although they are clearly not the same, the two should be equivalent at the point that universal primary education is achieved.

Because PRSPs are focused on short to medium term development strategies, targets are normally set at earlier dates than 2015. Therefore, comparisons often had to be made between “implied” targets by either using the extrapolation method or annualizing the growth rate. For example, for goals such as universal primary education, for which the target is to reach 100% by 2015, the comparison is made between the extrapolated target value in 2015 based on the current (or benchmark year) value and the target value set sometime before 2015. If the implied PRSP target in 2015 is more (less) than 100%, then the PRSP target was rated more (less) ambitious than the MDG target. For goals such as child mortality, for which the target is to reduce the level by a certain proportion – in this case, two-thirds – in 25 years (between 1990 and 2015), the implicit annual growth rate (-4.3%) was compared with the annual growth rate implied by the PRSP target, which is identified from the current (or benchmark year) value and the target year value. If a target was given, but no base year data were available in the PRSP to calculate the rate of change, we have used data from the WDI database, unless there is a substantial discrepancy between the PRSP data and the WDI database.If the implied PRSP annual growth rate is less (more) than -4.3%, the PRSP target was rated more (less) ambitious than the MDG target. All the implied rates are calculated assuming an exponential growth path, rather than a linear growth path.

How plausible are PRSP targets?

In any planning document there may be some tendency to pick overly ambitious or “stretch” targets to justify additional resources or give a greater sense of urgency to implementation of the proposed plan. On the other hand, risk averse planners might be inclined to select readily attainable targets, especially if they are likely to be held accountable for achieving or falling short of the goal. There is no clear means of distinguishing the two behaviors, but comparison with past results and with other externally given targets may yield some insights into the choices made.

Past rates of progress

A comparison was made of historical rates of progress on eight MDG indicators with the targeted rates in the PRSPs. To be included in the sample a country had to have given a baseline and target value for the indicator in the PRSP and to have data on the same indicator recorded in the World Bank’s WDI database for the period 1990 through 2000. In some cases, such as the poverty headcount ratio, these criteria severely limited the useful dataset. When the PRSP included more than one target year, the value in the nearest future year was used. Comparisons were made using the annualized exponential rate of change.

The results are summarized in Table 2. PRSPs for which the proposed target implies a faster rate of change than observed in the historical period were classified as “more ambitious.” Countries for which the target rate of change exceeded the historical rate by more than 5 percentage points were classified as “more than 5% ambitious.”


For countries with comparable data, PRSP target growth rates tend to be more ambitious than the historic growth rates except for gender indicator and primary completion rate. For malnutrition and water, all countries for which data are available (8 and 9 respectively) set PRSP targets growth rates that are more ambitious than the historic growth rates. There are a number of PRSP targets that are significantly more ambitious compared to the historic trend. For example, 3 out of 5 countries set assisted delivery targets that are more than 5% ambitious compared to the past trend, and 2 out of 4 countries set poverty headcount targets that are more than 5% ambitious than the past trend. Particularly large differences between the two growth rate due to a very ambitious PRSP targets are found for the following countries/indicators:

  • Mali, assisted delivery -- 29.3 percentage point difference
  • Burkina Faso, gross female primary enrollment rate -- 14.9 percentage point difference
  • Burkina Faso, immunization -- 14.9 percentage point difference
  • Nicaragua, assisted delivery -- 10.9 percentage point difference
Progress toward MDG targets

The comparison of the target rates of progress revealed that in cases for which the PRSPs include a well defined target and comparable base year data are available, PRSP targets in several areas are at least as ambitious as the MDG target. For example, 13 out of 14 PRSPs set child malnutrition targets that are at least as ambitious as the MDG target; 16 out of 18 PRSPs set access to improved water targets that are at least as ambitious as the MDG targets; and 9 out of 13 PRSPs set net primary enrolment targets that are at least as ambitions as the MDG target. This does not tend to be the case for targeted rate of progress to reduce infant, child, and maternal mortality rates, where PRSPs often set less ambitious targets than the MDG target (see Figure 1).

Figure 1: Target comparison between MDGs and PRSPs

Comparisons between PRSPs over time

The earliest PRSPs were begun before the advent of the MDGs. Although targets similar to those of the MDGs have been under discussion for much of the previous decade, it could be expected that earlier PRSPs might be less well aligned with the MDGs. As the MDGs became better known, particularly after the release of the Millennium Declaration Road Map in September 2001, country teams preparing PRSPs may have been more inclined to adopt similar targets and indicators.

We have not done a full comparative analysis of the coverage of earlier and later PRSPs, but the proportion of countries that do not have enough information to make the comparisons of their targets with MDG targets (either because they do not use the indicator for PRSP monitoring or because a target or a baseline of the indicator is not determined) is much lower among the recent PRSPs (24%) compared to the earlier ones (39%), indicating a somewhat greater alignment in indicator selection.