World Development Report 2024: The Middle-Income Trap

What indicates that an economy is developing? Income per capita is the most commonly used metric to measure the pace of economic development. But measures of average income can differ greatly, depending on the measure. For example, the World Bank, other international organizations, and bilateral aid agencies use GDP per capita at market exchange rates for analysis and lending (box 1.1). On the other hand, the use of GDP per capita based on adjustments for purchasing power parity (PPP, which reflects the purchasing power of a consumer for goods and services) can yield different results.

A comparison of these two sets of measures for Türkiye and Chile illustrates the problem. According to World Bank estimates that use market exchange rates, Türkiye is a middle-income country, and Chile is a high-income country. Yet when adjusted for purchasing power, Türkiye’s GDP per capita is higher than Chile’s. The country for which both measures are defined as identical is the United States because purchasing power in any country is measured relative to what a dollar can buy in the United States. Türkiye’s GDP per capita relative to that of the United States is nearly 50 percent when adjusted for PPP but less than 15 percent using market exchange rates. For Chile, the numbers are, respectively, 40 percent and 20 percent.

Box 1.1 Misunderstanding through misclassification

The World Bank’s income classification method for grouping countries was first presented in the 1978 World Development Report.a It introduced groupings of “low-income” and “middle-income” countries using a threshold of US$250 gross national income (GNI) per capita between the groups. The low-income threshold was set in keeping with the guidelines for procurement of goods and services for civil works projects for countries eligible for assistance from the International Development Association (IDA), the organization in the World Bank Group that supports the world’s least developed countries. Specifically, the threshold was based on the “civil works preference” operational guideline for IDA countries.

The process of setting thresholds for income per capita began with finding a “stable relationship” between a summary measure of well-being such as poverty incidence and infant mortality, on the one hand, and economic variables, including GNI per capita estimated using the World Bank’s Atlas method, on the other.b Based on such a relationship and the annual availability of the World Bank’s resources, the original income per capita thresholds were established.c They were last updated in 1989, using GNI per capita valued annually in US dollars based on a three-year average exchange rate and were expanded to four categories:

• Low-income. The low-income threshold was officially set in 1988, still based on the value of the IDA’s “civil works preference” and updated for inflation.d

• Lower-middle-income. The lower-middle-income threshold is based on the operational guidelines cutoff for determining access to 17-year repayment terms for loans through the World Bank Group’s International Bank for Reconstruction and Development (IBRD), although these terms are no longer available. It appears to have first been introduced in the 1983 edition of the World Development Report.e

• Upper-middle-income. The upper-middle-income threshold is the range between lower-middle-income and high-income.

• High-income. The high-income threshold does not relate to a cutoff derived from the operational guidelines, but was set at GNI per capita of US$6,000 in 1987 prices in a paper presented to the World Bank’s Board of Executive Directors in January 1989, which also reconfirmed the low- and lower--middle-income threshold levels.f The US$6,000 level has been updated over time for what is called “international inflation,” defined as the average inflation rates of Japan, the United Kingdom, the United States, and the euro area. The choice of the high-income threshold was made to address anomalies in the classification of high-income and industrialized economies used in the World Bank’s World Development Indicators prior to that point.

Under this current classification method, Zambia (with income per capita of US$1,170) and Bulgaria (with income per capita of US$13,250) are both middle-income economies. But few people would disagree with the observation that these countries have had vastly different development experiences and face vastly different growth challenges and trajectories. Yet these income classifications continue to be used widely in the development discourseg and in analyses of economic growth.h Moreover, although the World Bank may not use the income classifications for operational or lending purposes,i other international organizations and bilateral aid agencies do. Given its widespread use, many economists have called for a revision of the current income classification system.j The proposals include:

• Reclassifying levels based on fiscal capacity. Ravallion (2009) argues that levels of development should be assessed based on countries’ internal capacities for redistribution (through taxes) in favor of their poorest citizens. Similarly, Ceriani and Verme (2014) propose a measure of a country’s capacity to reduce its own poverty levels and show how these tools can be used to guide budget or aid allocations.

• Reflecting the multidimensional nature of development. Sumner and Vázquez (2012) use a set of indicators covering definitions of development from four conceptual frameworks (development as structural transformation; development as human development; development as democratic participation and good governance; and development as sustainability) to identify five types of developing countries. Similarly, Nielsen (2011) suggests the need to introduce a development taxonomy to classify countries based on a variety of existing development proxies (the Human Development Index, lifetime income measure, and so on) rather than income levels. Even the World Bank Group Strategy adopted in 2013 recognizes the need for an approach that pays more attention to the multiple facets and fragility across the development spectrum.k

b. For more on the Atlas method, see World Bank Atlas Method: Detailed Methodology (Data Help Desk), Data, World Bank, Washington, DC, https://datahelpdesk.worldbank.org/knowledgebase/articles/378832-what-is-the-world-bank-atlas-method.

i. Operationally, borrower countries are distinguished by their lending category within the World Bank: IDA-only; blend (both IDA and IBRD); and IBRD-only. IDA provides countries with the most difficulty borrowing externally with grants and concessional loans. IBRD offers nonconcessional loans to countries that it finds creditworthy. Because IBRD terms are “harder” (more market-based) than those of IDA, IBRD borrowers tend to be perceived as more developed than IDA borrowers. But graduating from low-income status is not the same as graduating from IDA. Eligibility for IDA benchmarks a country’s income against different thresholds, while graduation from IDA takes into account factors other than income. See Dolan (2016) for more information.

PPP adjustments have been criticized for their inability to reflect the complexity and diversity of economic production and capabilities in individual economies. The adjustments do not consider quality or productivity differences among countries in the production of tradable as well as nontradable goods and services, including infrastructure, health care, and education.

However, recent assessments of PPPs appear to be highly correlated with economic activity. A comparison of countries using PPP adjustments could produce a better understanding of the distance to the technology frontier than gross national income. However, income or GDP per capita does not reflect the wide array of growth challenges that countries face. Two high-income countries provide an example. In 2022, Qatar’s GDP per capita was US$88,046, driven mostly by exports of hydrocarbons from its abundant reserves of oil and natural gas. The same year, Denmark’s GDP per capita was US$66,983, and its services sector employs about 80 percent of labor. Meanwhile, cthe countries vary significantly in their levels of technical sophistication; in 2021, Denmark was ranked twenty-fourth in the Economic Complexity Index and Qatar eighty-second.