World Development Indicators 2004 (World Development Indicators)

Rating:
Summary: Almost Perfect
Review: In 1997, with little fanfare, the World Bank replaced the statistical appendix of its venerable annual publication, World Development Report, as well as the Bank's World Tables and Social Indicators of Development, with a new publication, the World Development Indicators (WDI). The WDI became "...the World Bank's principal statistical survey of world development." (WDI 1997, p. vii) It is a notable improvement over the discontinued publications, eliminating most but not all of weaknesses that previously frustrated both the regular and casual user.

Most importantly, the WDI provides more data (947 data sets in eighty-three tables compared to the 226 data sets in the eighteen tables in the 1996 World Development Report) on more countries. Detailed coverage is provided for 152 countries with more limited data on an additional 56 countries or other political entities. In addition to data gathered by the World Bank and associated organizations, the WDI includes developmental data from a wide variety of international, governmental, non-governmental, and private organizations. Extended time series for all of the data in the paper WDI are available on the associated WDI CD-ROM.

Each of the six sections in the WDI begins with a three to thirteen page introduction before presenting the relevant data sets. After the wide ranging World View chapter (6 tables and 76 data sets); the WDI looks at population, poverty, education and health in People, the second chapter (14 tables and 199 data sets). Agriculture, pollution, biodiversity, and environmental sustainability data are given in the Environment chapter (15 tables and 161 data sets). The remaining three chapters focus on more traditional measures of economic development. The chapter entitled Economy contains data on output, international trade, government spending and finance, as well as external debt (17 tables and 189 data sets). This is followed by the States and Markets chapter which contains data on a wide range of issues including stock market and other investments, tax policies, defense spending, transport, power and communications, and science (12 tables and 138 data sets). The sixth and final chapter, Global Links, contains data on trade blocs and tariffs, global financial flows, international labor migration, and foreign aid (14 tables and 184 data sets). The tables follow a cross sectional format to facilitate inter-country comparisons and many show data from two or more years so that trends can be observed.

The most useful innovation in the WDI is its treatment of definitions and sources. In its predecessor publications, each table's definitions and sources were generally obscure, laconic and hidden in the back of the book. A long distance - mentally if not physically - from the data. In the WDI, each of the eighty-three tables is immediately followed by a detailed discussion of the meaning, the strengths and, most importantly, the weaknesses of the data. Also included is the source of each data set which facilitates further research. This careful attention to keeping the reader informed about the source and quality of data should produce more light, if not less heat, when the statistics of development are used in research, education or in support of policy.

But the World Bank should go further and make an explicit estimate of the possible measurement error for each data set. As Morgenstern argued over fifty years ago in his seminal work On the Accuracy of Economic Observations (1950, second edition in 1963, Princeton University Press), it is much more efficient for the data producer to provide explicit estimates of data accuracy than for the data consumer to attempt the same exercise. Errors in published data can arise from a variety of causes including definitional weaknesses, simple recording errors, improper aggregation, time factors, and the deliberate falsification of information. Without an estimate of the scale of the measurement error of each data set, it is difficult to use this data correctly. For example, according to Table 1.1 in the WDI, entitled "Size of the Economy" (pp. 14-17), GNI per capita on a PPP basis in 2002 was $5,220 for Venezuela and $6,150 for Colombia. How confident are we that the good people of Colombia are actually better off then their neighbors in Venezuela? In other words, is this $930 gap (approximately 15% of the Colombian GNI estimate) significant or not? If the gap is significant then there might be policy implications. If the gap is less than the measurement error for this data then any attempt to draw conclusions from the income gap would be wrong.

Unfortunately, while the WDI often provides general warnings about data quality, there is no attempt to provide an estimate of possible measurement error for any of the data series. As a result, even the careful reader is never sure whether the comparisons of different levels or rates are meaningful. That estimates of data accuracy are difficult, arbitrary, and, to some extent, dependent upon the purpose to which the data is to be used makes it even more important that the World Bank make the effort.

The other serious weakness in the WDI is the limited and confusing coverage of Taipei, China (Taiwan). In eight of the tables, data on the Taiwanese economy is either listed separately or as part of the data for the People's Republic of China. In the other seventy-five tables, Taiwan is not included. It is a serious failing to exclude a $529 billion economy (GDP-PPP), the 17th largest in the world, an economy roughly the same size as Australia, because of political considerations. This exclusion not only limits the usefulness of the WDI but also probably weakens the Bank's efforts to encourage developing country's statisticians to withstand pressure from their own politicians.

Despite these weaknesses, the World Development Indicators is a extremely valuable resource for anyone with a professional, academic, or general interest in the social and economic situation of the world.

Morgenstern, Oskar (1963, first edition 1950) On the Accuracy of Economic Observations (2nd ed), Princeton: Princeton University Press.

World Bank (1996) World Development Report: From Plan to Market, Washington D.C.: World Bank.