Author Archives: admin

2021 September Update: The new update of the Barro-Lee data set is now available.

Barro-Lee Estimates of Educational Attainment for the Population Aged 15-64 from 1950 to 2015

This is the latest updated version of the Barro-Lee dataset reported in Barro and Lee (2013). Dr. Hanol Lee, an associate professor at Southwestern University of Finance and Economics, has collaborated on the project.

The main aim of this new version is to construct estimates of educational attainment for the population between 15 and 64 years old for the year of 2015. The estimates are disaggregated by gender and by 10-year age group, whereas those in the original dataset were disaggregated by 5-year age group. This is due to the limited availability of disaggregated statistics in the newly complied census/survey data. The methodology to construct the estimates for the year of 2015 is as follows:

  • We collect new census/survey figures on educational attainment of the population for 2015 (or a near year). They include a total of 96 observations among the sample of 146 countries. The census/survey data are compiled from the United Nations Educational, Scientific and Cultural Organization (UNESCO) Institute for Statistics, United Nations (UN) demographic yearbook, Eurostat, Organisation for Economic Co-operation and Development (OECD), and national statistics agencies. The following table provides detailed information on the source of the observations.
  • The data availability (i.e., structure of the census/survey figures) varies by the source. The Eurostat and UN demographic yearbook report the distribution of educational attainment for the population between 15 and 64, which can be disaggregated by 10-year age groups. On the other hand, OECD database presents the distribution of educational attainment for the population aged 25-64, disaggregated by age groups. UNESCO reports only aggregate statistics for the population aged 25 and over. Most census/survey figures report the distribution of educational attainment by schooling level following the UNESCO’s “International Standard Classification of Education” (ISCED)—that is, four broad-level classifications—no formal education, primary, secondary, and tertiary education—and then the breakdown of primary, secondary, and tertiary education into subcategories. But Eurostat and OECD provide numbers only a combination of several categories.
  • We fill in missing observations in all 20 categories (5 age groups × 4 education levels) for each country for the year of 2015 to construct complete estimates of educational attainment for the population aged 15-64 from 1950 to 2015, disaggregated by 10-year age groups (15-24, 25-34, 35-44, 45-54 and 55-64). We first fill in the missing observations at four broad classifications by adopting the ‘forward extrapolation method’ which uses the census/survey figures in 2010 or an earlier year as a benchmark. We then apply estimates of completion ratios to break down the estimate in broad categories and fill in the observations that are missing in subcategories.
  • We use the updated information on school enrollment ratios, drop-out ratio, and population structure for the estimation of educational attainment of the population.

We combine these updated estimates for 2015 with the original Barro-Lee data set from 1950 to 2010 and present a new data set of educational attainment for the population aged 15-64 from 1950 to 2015, disaggregated by 10-year age groups (15-24, 25-34, 35-44, 45-54 and 55-64).

  • We also collect 7 new censuses on educational attainment of the population for earlier years from the UN Demographic Yearbook. They consist of two countries in 2005 (Egypt and New Zealand), and five countries in 2010 (Argentina, Armenia, Kazakhstan, Latvia, and Saudi Arabia). In addition, we collect 3 new observations for Iceland in 2000, 2005 and 2010 from Eurostat. Accordingly, we revise the original Barro-Lee estimates for these countries through backward extrapolation incorporating these newly available censuses. Information from these censuses is also utilized for the estimation of the figure for 2015 through forward extrapolation, when necessary.

Note: The dataset below includes the revised estimates for USA in 2015. The new estimates are based on the US Current Population Survey (CPS) in 2015. The earlier version (released on September 2021) included the US estimates in 2015 based on the OECD source. The revised data for the USA is here.

Data

Download full data set (146 countries) for the Population aged 15-64, aged 25-64 or by 10-year age group in 10-year interval (1950-2015) in xls, csv, or dta format.

Education Attainment for Population Aged 15-64

Total population

Excel CSV STATA

Female population

Excel CSV STATA

Male population

Excel CSV STATA

Education Attainment for Population Aged 25-64

Total population

Excel CSV STATA

Female population

Excel CSV STATA

Male population

Excel CSV STATA

Education Attainment by Age Group

Total population

Excel CSV STATA

Female population

Excel CSV STATA

Male population

Excel CSV STATA

2018 6 June Update

In the version 2.1, there were minor mistakes in the estimates for 10 countries (Albania, Bolivia, Cameroon, Dominican Rep., Egypt, Fiji, Ghana, Kenya, South Africa, Sweden). The data set of version 2.2, which reflects the corrections, replaces the version 2.1 and is available at the data page.

2016 4 Feb. Update

In the version 2.0, there was a mistake in the estimates for China. The data set of version 2.0 is changed accordingly and available as version 2.1 at the data page.

2014 30 Aug. Update

In the version 2.0, there was a mistake in the estimates for Finland in 2005 and 2010. The data set of version 2.0 is changed accordingly and available at the data page. We thank Aleksi Kalenius for pointing out this mistake.

2014 30 June Update

The new version (2.0) is now available.

This version of the Barro-Lee data set has made several important changes to the earlier version (1.3) reported in Barro and Lee (2013). Hanol Lee, a Ph.D. Student in Economics at Korea University, has provided excellent research assistance.

The changes affected various countries in the sample. The major changes are as follows:

  • We have updated estimates of educational attainment by using the recently available UNESCO census data. They include 89 census observations from the 86 countries, mostly for the years 2005 and 2010. The census data on educational attainment of the population by age and by gender are kindly provided by the UNESCO Institute for Statistics.
  • We have also collected the new census/survey observations from the UN demographic yearbook and national publications from statistics bureaus.
  • We have added 126 census/survey observations in total and used them as benchmark figures for the estimation of educational attainment. The following Table provides detailed information on these additional observations.
  • We have corrected minor errors in the UNESCO censuses, including Korea in 1990 and Mexico in 2000.
  • To maintain consistency with the observations from the other sources, we have reformulated the completion ratio of tertiary-level attainment for the census observations sourced from the UN Demographic yearbooks. The tertiary completion ratio is defined as the percentage of the number of people who have attained schooling at ISCED 5A or 6 in the number of people who have attained schooling at ISCED 5A, 5B or 6. The people who have ever enrolled in the 4th year of colleges or universities belong to the complete category.
  • We have used the updated data on enrollment ratios for total and female school-aged population at each education level for 2005 and 2010.
  • In this version, we have checked more carefully the consistency between the estimates of the distribution of educational attainment among total, female and male population by age group. We report the data for male population by age group together with the ones for total and female population.
  • We are currently constructing estimates of historical educational attainment from 1870 to 1945 in five-year intervals. For the estimation, in this version (2.0), we have used newly compiled data on school duration data and school enrollment ratios for the years before 1950, and adopted a modified backward extrapolation technique. Accordingly, there are some changes in the estimates for the older-age population after 1950 (that is, those derived from the estimates for school-age population before 1950). We will report historical data on school duration and school enrollment ratios, as well as a new data set of historical educational attainment, with new estimation procedures, at the next opportunity.
  • The version (1.3) released on April 2013 is also available.

2013 09 April Update

The new version (1.3) is now available!

This version (1.3) of the Barro-Lee data set has made several changes to the earlier version (1.2). The changes affected various countries in the sample, but they are not significant in magnitude for most cases.

The changes are:

  • We have evaluated more carefully the accuracy of the forward-flow and backward- flow estimates of educational attainment for the total and female population in the age group, 15-19 and 20-24. We have made corrections to those estimates that showed unrealistic fluctuations over time, which are mainly due to measurement errors in school enrollment rates. The new estimates display a smoother trend in each category of educational attainment and average years of schooling among the younger cohorts than the previous estimates.
  • We have checked the accuracy of the estimates of educational attainment among male population by age group, which are constructed from the estimates among total and female population by age group. We have made corrections to the estimates that showed unrealistic fluctuations over time.
  • We have changed the backward extrapolation method. We continue to apply the backward extrapolation procedure to fill in missing data for each age group by using the attainment of the older age group from the succeeding period. But it turns out that the backward-flow extrapolated estimates for the age group below 65 years old are not often accurate if we use the attainment of the age group over 65 years old from the succeeding period as benchmark data when they are not from census. So, we decide not to use the backward extrapolation method for this case. Instead, we have estimated the attainment of the age group, 60-64 by using the group-average age-specific profile for the same age group constructed using the available data of the countries in advanced countries or developing countries, and then used the estimates as benchmark data for the backward extrapolation for the age group in the preceding period.
  • We have used recently available UNESCO data of secondary and tertiary enrollment rates for 2010 to estimate missing attainment data. This has affected secondary and tertiary educational attainment among the population in the age group, 15-19 and 20-24.

2011 04 Sept. Updat

The new version (1.2) is now available!

This version (1.2) improves on the earlier version (1.1) by incorporating recently available census observations and correcting inaccurate estimates of completion ratios.

  • The new census data are now available through UNESCO and national statistical offices.
  • The census data for the following countries are available through UNESCO: Canada, Fiji, Ghana, Guyana, Iran (Islamic Republic of), Maldives, Namibia, Papua New Guinea, Tonga and United Arab Emirates; while the census data for the following countries are available through their respective statistical offices: Macro SAR, Cote d’Ivoire, Guatemala, Italy, Russian Federation, Slovenia, Sri Lanka and United States.
  • Some countries have extremely low or high county specific primary/secondary completion ratios. As these imply unusual trends in the completion ratios, we replaced them by the regional specific primary/secondary completion ratios. These countries include: Afghanistan, Algeria, Austria, Botswana, Democratic Republic of the Congo (only F), Hungary, India, Malta, Myanmar, New Zealand, Portugal (only MF), Sierre Leone, United Kingdom and Zambia.
  • A more detailed list of these revisions is available here.

History

Many observers have emphasized the crucial importance of human capital, particularly as attained through education, to economic progress (Lucas, 1988, Barro, 1991 and Mankiw, Romer and Weil, 1992). An abundance of well-educated people goes along with a high level of labor productivity. It also implies larger numbers of more skilled workers and greater ability to absorb advanced technology from developed countries. The level and distribution of educational attainment also have impact on social outcomes, such as child mortality, fertility, education of children, and income distribution.

There have been a number of attempts to measure educational attainment across countries to quantify the relationship between it and economic and social outcome variables. Earlier empirical studies used school enrollment ratios or literacy rates. But although widely available, these data do not adequately measure the aggregate stock of human capital available contemporaneously as an input to production.

Earlier versions of the Barro-Lee Data Set (1993, 1996, and 2001) filled this data gap by constructing measures of educational attainment for a broad group of countries.

The earlier versions employed perpetual inventory method using census/survey observations on the educational attainment of the adult population group over age 15 or over age 25 as benchmark stocks and new school entrants as flows that added to the stocks with an appropriate time lag. The flow estimates were estimated using information on school-enrollment ratios and population structure over time.

As new data becomes available, the data set is updated and expanded. Barro-Lee (1993) provides educational attainment estimates for 129 countries for 1960–1985. This new Data Set provides complete estimates for 146 countries for the period 1950-2010.

Methodology

We fill in most of missing observations by forward and backward extrapolation of the census/survey observations on attainment. The estimation procedure extrapolates the census/survey observations on attainment by 5-year age groups at five-year intervals fill in missing observations with an appropriate time lag.

We assume that an individual’s educational attainment remains unchanged from age 25 to 64 and that mortality is uniform across all individuals, regardless of educational attainment. Hence, for age groups between 25 and 64, we fill the missing attainment data using the attainment of the younger age group from the previous period (forward) as benchmark or the attainment of the older age group from the succeeding period (backward).

Since direct backward or forward extrapolation is not applicable for the two youngest age groups (age 15-19 and 20-24), we use attainment and enrollment data to estimate missing attainment data. We assume that the change in enrollment leads to a proportional change in attainment over time with time lag. Hence, for these age groups, we use estimates for the same age group from the previous (or in the next) period as benchmark and adjust this benchmark figure by the change in enrollment over time or the enrollment adjustment factor.

For older age groups (age 65 and above), we distinguish between a less-educated population (uneducated and people who have reached the primary level) and a more-educated population (reached at least secondary schooling). We estimated the survival rates for the old population by education levels using available censuses by age group and found that more educated people have lower mortality rates. We apply our survival ratio estimates to adjust the backward or forward estimate for mortality rate differences between less-educated and more-educated individuals.

After estimating school attainment at four broad levels of schooling: no school, some primary, some secondary, and some higher, we break down the three levels of schooling into incomplete and complete education by using estimates of completion ratios.

For a more detailed discussion about the estimation methodology, see Barro-Lee (2013).