Developing a comprehensive time series of GDP per capita for 210 countries from 1950 to 2015
Published in Population Health Metrics, July 2012
New data published in the study “Developing a comprehensive time series of GDP per capita for 210 countries from 1950 to 2015” track gross domestic product (GDP) over six decades. Researchers from IHME used models to fill in gaps in time and across 210 countries for existing GDP datasets and created two new GDP time series.
Income has been extensively studied and utilized as a determinant of health, but while there are several datasets commonly used to study GDP per capita, there are no time series that are complete for the years between 1950 and 2015 for the 210 countries for which data exist. This study set out to achieve two goals: 1) to produce a complete time series for each existing GDP dataset for all countries and years between 1950 and 2015 and 2) to create a new United States dollar and international dollar series based on the completed source series for all countries and years between 1950 and 2015.
This study produced nine complete GDP per capita time series from 1950 to 2015 for 210 countries. The authors also explored whether the different series yield different results regarding the relationship between income per capita and under-5 and adult mortality. They found notable interseries variation, indicating that the choice of series will significantly affect conclusions drawn about the relationship between GDP and mortality. However, using the completed data series removes this problem, suggesting that the development of the complete income series is important not only because it allows for more comprehensive modeling, but also because using the complete income series moderates variation in quantitative inferences that can be drawn.
Researchers collected GDP per capita estimates expressed in either constant US dollar or international dollar terms from all seven major sources, including the University of Pennsylvania, World Bank, United Nations Statistics Division, International Monetary Fund, and Angus Maddison’s research homepage at the University of Groningen Department of Economics, three of which were expressed in US dollars and four of which were expressed in international dollars. The researchers applied several stages of models, including ordinary least-squares regressions and mixed effects models, to complete each of the seven source series from 1950 to 2015. The three US dollar and four international dollar series were each averaged to produce two new GDP per capita series for all countries from 1950 to 2015.
The authors note that these series are the most comprehensive GDP per capita series currently available and offer researchers diverse options to serve different analytical purposes. Extending the existing data to missing countries and years should facilitate population health analyses and reduce biases that arise from using series with missing country-years. The strength of the IHME data series is that they reduce the bias that may result from using one source’s series for analyses, and the authors encourage researchers to test the sensitivity of their findings using an alternative completed time series, regardless of which one of the nine completed datasets they choose to use initially.
Citation: James SL, Gubbins P, Murray CJL, Gakidou E. Developing a comprehensive time series of GDP per capita for 210 countries from 1950 to 2015. Population Health Metrics. 2012; 10:12.