In column III we additionally look at the age composition of popula- tion. We add two variables, namely the percentage of population that is below 14 and the percentage of population that is between 14 and 64 years old. Note that the share of elderly people above 64 years cannot be simultaneously added as the three shares add up to one and are therefore collinear. Adding the age composition hardly affects the population elas- ticity, which remains close to unity. Neither the affluence nor the tech- nology variables are much affected in either this or consecutive estimations and are therefore not further discussed. A higher percentage of the.
