Databases and Development Economics

Takashi Kurosaki


A summer extension lecture course was held at the Institute of Developing Economies, ( hereafter called Ajiken ), Tokyo on the 27th and 28th of July, 1998 under the title of "Development Economics:How to Handle and Utilize Data." Each lecture was well attended, indicating a surge of interest in the subject. The present article introduces recent trends in the use of databases in development economics by reviewing these lectures and deliberates upon the implications such trends may have on the Asian Historical Statistics Database Project ( AHSP ).

1. Statistical Data for Analyses of Economic Development

The summer course of Ajiken was comprised of six lectures;" National Accounts and Production," "Trade," "Multi-national Enterprises ( MNE ) and Foreign Direct Investment ( FDI )," "Use of Demographic Data," "Agriculture and Household Economies," and "Welfare and Environment."( See the November 1998 issue of Ajiken World Trends for the contents of each lecture ). Among these lectures, the author was in charge of " Agriculture and Household Economies." The six lectures shared a practical approach with the emphasis on how to obtain data on developing economies and how to utilize them in analyzing development issues. This is also the approach by the AHSP.

Among the items covered by the course, national accounts, census of ( manufacturing ) industries, trade statistics, demographic census, and agriculture census and production statistics: there are used within the AHSP. Some of the data in developing countries were once less accurate than those in developed countries and required extra care when using them. But today the level of accuracy has been improved and various indexes have been prepared and published in almost all Asian countries. Consequently, there seems to be a consensus among area specialists in Ajiken that existing statistics can be utilized to a considerable degree. There are, however, some statistics such as trade statistics that require special care even today because of changes in categorization.

In contrast to this, databases for earlier years are extremely limited both in quantity and quality. Differences in the initial conditions are the hot issue in recent models of endogenous growth theories, but statistics that indicate such differences are hard to find, or not so reliable even if they are available. The Ajiken course once again reassured us that the AHSP could contribute to the solution of this problem.

Statistics on " MNE and FDI "and " Welfare and Environment "are not included in the AHSP. As indicated by the fact that each of them was allocated an independent lecture, they are key factors that will affect the future of developing economies, both theoretically and practically.

The Ajiken summer course pointed out that, first, the existing data are too limited even for fact-finding studies, and, second there are wide differences in definition and concepts, which make it difficult to compare statistics across national borders. Extreme caution must be used when handling these existing statistics. Furthermore, since interest in these topics emerged relatively recently, there are considerable difficulties in acquiring such information for a long historical period.

Therefore, the AHSP has not adopted these two as independent subjects. However, as a second best measure, it is possible to deal with information related to these topics within the present framework. For example, it would be possible to incorporate FDI data in balance of payment statistics, to incorporate data about the soil quality in agricultural production statistics, or to expand the coverage of social statistics.

2. Use of Microeconomic Data from Large-Scale Sample Surveys

Demographic and agricultural statistics in aggregate forms have been collected from the oldest period, and consequently they are given emphasis in the AHSP. Recently, however, quantitative analyses in these areas have rapidly become more micro-oriented, making the distance from the AHSP greater.

In demography, many researchers use Demographic and Health Surveys ( DHS ), carried out by Macro International, an American survey organization, with the financial support of USAID. DHS started in 1984 and continues today, covering about 60 developing countries with a sample size from 3,000 to 20,000. Household-level data are available with full access. World Fertility Surveys ( WFS ) provide similar data for surveys conducted during the 1970s. Utilizing these databases is now a common practice to analyze birth, death, immigration, etc. at the household level.

In the analyses of rural economies, large microeconomic data sets have been also used intensively. Particularly, panel data that cover both production and consumption of agricultural households over several years have attracted much attention. The well-known example is the panel data compiled by the International Crops Research Institute for the Semi-Arid Tropics ( ICRISAT ) in the Deccan region of India, collected over a period of ten years. It is not an exaggeration to state that the ICRISAT data fostered new development economics, in which theory and empirical tests are closely combined to analyze the microeconomics of intra-household relations, household dynamics, and inter-household relations.

Lately, the World Bank's Living Standard Measurement Surveys ( LSMS ) have been attracting great attention. This project was launched in 1979 and is still in progress. Comprehensive information about living standards, poverty, inequalities etc. is gathered in line with the system of national accounts (SNA) so that the data facilitates international comparison. The project attempts to compile information into panel data as much as possible. So far, more than twenty countries have been covered. The sample size is from several thousand to over ten thousand per country. Detailed data of high quality awaits access not yet fully explored - the LSMS data archive is a gold mine for researchers who are interested in microeconomic analyses of developing economies.

An LSMS survey is usually conducted with the collaboration of government statistics agencies in each country. In such joint efforts, know-how about sample surveys has been transferred and accumulated in countries where large-scale sample surveys had not been conducted previously. The high reliability of the LSMA database is a direct result of such joint research while, as an indirect result, macroeconomic statistics such as GNP and its consumption statistics are expected to improve in quality.

It may not be realistic to bring these micro databases into the AHSP, which basically aims at macro statistics. It, however, would be of significance to introduce these data sets in the commentaries of statistics or in reports by the statistics institution team. In addition, it would be promising to examine the consistency between macro statistics such as GNP, agricultural production, and demographic statistics, and micro databases from large-scale sample surveys.

3. Access to Data

One characteristic of the new databases introduced so far is that they can be easily accessed through the Internet or CD-ROMs and that they can be dealt with on personal computers. The World Bank's LSMS project, for instance, makes it a rule to open the database at the micro level. If you access it on the web site ( http://www.worldbank.org/lsms/lsmshome.html ), you can easily see the availability of data, details about the survey, and the access policy in the latest version. Some countries allow you to download the data on the spot. As for DHS, almost all the micro data can be downloaded through the Internet ( http://www.macroint.com/dhs )

There are many other useful databases from developing countries, such as family life surveys by the RAND Corporation. As a macro economic database, Penn-World Table is especially important. See for example, http://chaka.sscnet.ucla.edu/dev.data/data.htm for a convenient link site for data from developing countries. These web sites provide us with various ideas about the way we might make public the final products of the AHSP.

(Institute of Economic Research, Hitotsubashi University)
Translated by WOW LI.