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).