Southern African Regional Poverty Network (SARPN) SARPN thematic photo
Country analysis > Malawi Last update: 2008-12-17  
leftnavspacer
Search





 Related documents

USAID
Economic Performance Assessment: Malawi

Bruce Bolnick, Rose Mary Garcia, Alex Greenbaum, Maureen Hinman, Gertrude Mlachila

Nathan Associates

April 2005

SARPN acknowledges the Development Experience Clearinghouse website as the source of this report: www.dec.org
[Download complete version - 482Kb ~ 3 min (41 pages)]     [ Share with a friend  ]

Introduction

This paper is one of a series of Economic Performance Assessments (EPAs) prepared on behalf of the EGAT Bureau to provide USAID missions and regional bureaus with a concise analysis of selected economic growth (EG) performance indicators for particular host countries. The aim is to help USAID missions gain a clear picture of the host economy, as an input into the identification of possible strategic priorities for EG program interventions. The review uses international comparisons (“benchmarking”) to highlight major constraints, trends, and opportunities in areas such as macroeconomic management, trade policy, financial markets, the legal and regulatory environment, agricultural development, and others enumerated below. The analysis draws on the latest data from USAID’s internal Economic and Social Database (ESDB)1 and from readily accessible public information sources.

The approach used here is analogous to examining an automobile dashboard to see which gauges are signaling problems. A blinking light sometimes has obvious implications—such as the need to fill the fuel tank when the indicator shows that the tank is low. In other cases, it is necessary to have a mechanic probe more deeply to assess the source of the trouble and discern the best course of action.2 The EPA, similarly, is based on an examination of key economic and social indicators. For some of the issues where indicator lights are blinking, a detailed study may be needed to investigate the problems more fully and identify appropriate programmatic interventions.

Analytical framework

The analysis is organized around two interrelated and mutually supportive goals: transformational growth and poverty reduction.3 Rapid and broad-based growth is the most powerful instrument for poverty reduction. At the same time, measures to invest in human capital, reduce poverty, and lessen inequality help to underpin rapid and sustainable growth. These interactions create the potential for a virtuous cycle of economic transformation and human development.

Transformational growth requires a high level of investment and rising productivity. This is achieved by establishing a strong enabling environment for private sector development, involving multiple elements: macroeconomic stability; a sound legal and regulatory system, including secure contract and property rights; effective control of corruption; a sound and efficient financial system; openness to trade and investment; sustainable debt management; investment in education, health, and workforce skills; infrastructure development; and sustainable use of natural resources.

The impact of growth on poverty depends on policies and programs that create opportunities and build capabilities for the poor. We call this the pro-poor growth environment.4 Here, too, many elements are involved, including effective education and health systems; a strong commitment to fighting HIV/AIDS; policies facilitating job creation; agricultural development (in countries where the poor depend predominantly on farming); dismantling barriers to micro and small enterprise development; and progress toward gender equity.

Criteria for selecting indicators

The scope of the paper is constrained by the availability of suitable indicators. Indicators have been chosen to balance the need for broad coverage and diagnostic value, on the one hand, and the need of brevity and clarity, on the other. The analysis covers 15 EG-related topics, and just over 100 variables. For the sake of brevity, the write-up highlights issues for which the “dashboard lights” appear to be signaling serious problems, which suggest possible strategic priorities for USAID intervention. An accompanying Data Supplement provides a full list of indicators, along with the complete Malawi data set, including data for the benchmark comparisons, and technical notes for every indicator.

For each topic, the analysis begins with a screening of primary performance indicators. These “level I” indicators are selected to answer the question: Is the country performing well or not in this area? The set of primary indicators also includes descriptive variables such as per capita income, the poverty head count, and the age dependency rate.

In areas of weak performance, the analysis proceeds to review a limited set of diagnostic supporting indicators. These “level II” indicators provide more details about the problem or shed light on why the primary indicators may be weak. For example, if economic growth is poor, one can examine data on investment and productivity as diagnostic indicators. If a country performs poorly on educational achievement, as measured by the youth literacy rate, one can examine determinants such as expenditure on primary education, and the pupil-teacher ratio.5

Particular indicators have been selected on the basis of several criteria. Each indicator must be accessible through USAID’s Economic and Social Database or convenient internet sources. The indicators must be available for a large number of countries, including most USAID client states. Each one must be sufficiently timely to support an assessment of country performance that is suitable for strategic planning purposes. Data quality is another paramount consideration. For example, subjective survey responses are used only when actual measurements are not available. Aside from a few descriptive variables, the indicators must also be useful for diagnostic purposes. Preference is given to measures that are widely used, such as Millennium Development Goal indicators, or evaluation data used by the Millennium Challenge Corporation. Finally, redundancy is minimized. If two indicators provide similar information, one is selected, with preference to variables that are simplest to understand. For example, both the Gini coefficient and the share of income accruing to the poorest 20% of households can be used to gauge income inequality. We use the income share because it is simpler, and more sensitive to changes.

Benchmarking methodology

Comparative benchmarking is the main tool used to evaluate each indicator. The analysis draws on several criteria, rather than a single mechanical rule. The starting point is a comparison of performance in Malawi relative to the average for countries in the same income group and region — in this case, low-income countries in sub-Saharan Africa (hereafter “LIC-Africa”).6 For added perspective, three other comparisons are examined: (1) the global average for this income group; (2) respective values for two comparator countries selected by the Malawi mission (Uganda and Mozambique); and (3) the average for the five best and five worst performing countries globally. Most comparisons are framed in terms of values for the latest year of data from available sources; in cases where year-to-year fluctuations are large, five-year averages are used. Five-year trends are also taken into account if they shed light on the performance assessment.7

For selected variables, a second source of benchmark values uses statistical regression analysis to establish an expected value for the indicator, controlling for income and regional effects.8 This approach has three advantages. First, the benchmark is customized to Malawi’s specific level of income. Second, the comparison does not depend on the exact choice of reference group. Third, the methodology allows one to quantify the margin of error and establish a “normal band” for a country with Malawi’s characteristics. An observed value falling outside this band on the side of poor performance signals a serious problem.9

Finally, where relevant, Malawi’s performance is weighed against absolute standards. For example, Malawi’s inflation rate averaged 20% over the past five years. Regardless of the regional comparisons or regression results, this is a sign of serious economic mismanagement.

The results of this exercise must be interpreted with caution. No analysis of this sort can provide mechanical or definitive answers to questions about strategic priorities. For some topics, such as macroeconomic policy, it is easy to find fairly clear diagnostic indicators. For others, such as the quality of economic infrastructure, international statistics tell a very incomplete story. The aim is to identify signs of serious economic growth problems based on a systematic review of a variety of indicators, subject to the limits of data availability and quality, and thereby provide analytical insight into possible priorities for USAID interventions. On-the-ground knowledge and further indepth studies are required to supplement this broad-strokes analysis.

The remainder of this report discusses the most important results of the diagnostic analysis. The review is presented in three sections: Overview of the Economy; Private Sector Enabling Environment; and Pro-Poor Growth Environment. Table 1-1 summarizes the topic coverage.

Table 1
Topic Coverage

Overview of the Economy Private Sector Enabling Environment Pro-Poor Policy Environment
  • Growth Performance
  • Poverty and Inequality
  • Economic Structure
  • Demographic and Environmental Conditions
  • Gender
  • Fiscal and Monetary Policy
  • Business Environment
  • Financial sector
  • External sector
  • Economic Infrastructure
  • Science and Technology
  • Health
  • Education
  • Employment and Workforce
  • Agriculture



Footnotes:

  1. The ESDB is accessible through the USAID intranet. It is compiled and maintained by the Development Information Service (DIS), under PPC/CDIE.


  2. Sometimes, too, the problem is faulty wiring to the indicator—analogous here to faulty data.


  3. In USAID’s White Paper on U.S. Foreign Aid: Meeting the Challenges of the Twenty-first Century (January 2004), transformational growth is a central strategic objective, both for its innate importance as a development goal, and because growth is the most powerful engine for poverty reduction.


  4. A comprehensive poverty reduction strategy also requires programs to reduce the vulnerability of the poor to natural and economic shocks. This aspect is not covered in the template since the focus is on economic growth programs. Also, it is difficult to find meaningful and readily available indicators of vulnerability to use in the template.


  5. Deeper analysis of the topic using more detailed data (level III) is beyond the scope of papers in this series.


  6. Income groups as defined by the World Bank for 2004. For this study, the average is defined in terms of the mean; future studies will use the median instead, because the values are not distorted by outliers.


  7. The five-year trends are computed by fitting a log-linear regression line through the data points. The alternative of computing average growth from the end points produces aberrant results when one or both of those points diverges from the underlying trend.


  8. This is a cross-sectional OLS regression using data for all developing countries. For any indicator, Y, the regression equation takes the form: Y (or ln Y, as relevant) = a + b * ln PCI + c * Region + error – where PCI is per capita income in PPP$, and Region is a set of 0-1 dummy variables indicating the region in which each country is located. Once estimates are obtained for the parameters a, b and c, the predicted value for Malawi is computed by plugging in Malawi-specific values for PCI and Region. Where applicable, the regression also controls for population size and petroleum exports (as a percentage of GDP).


  9. This report uses a margin of error of 0.66 times the standard error of estimate (adjusted for heteroskedasticity, where appropriate). With this value, 25% of the observations should fall outside the normal range on the side of poor performance (and 25% on the side of good performance). Some regressions produce a very large standard error, giving a “normal band” that is too wide to provide a discerning test of good or bad performance.



Octoplus Information Solutions Top of page | Home | Contact SARPN | Feedback | Disclaimer Get Adobe Reader