Macroeconomic time series of human capital are computed for France ove
r a long period. The educational level is used to discriminate the qua
lity and the efficiency of different types of labour. Computations use
the perpetual-inventory method. In addition, we make use of vital sta
tistics concerning the population age and gender structure and the par
ticipation (activity) rates, of educational data concerning degrees at
tainments and years of schooling, and of surveys concerning the link b
etween wages and educational levels. It allows us to compute several i
ndicators of educational capital, investment and scrapping:the bachelo
r stock in the 15-65 aged population, the structure of the working pop
ulation, the <<stock of education>> in personyears of schooling. The
average years of schooling and the structure of the working population
according to years of schooling evolve gradually and are little distu
rbed by wars, but show two breakings around 1960 (baby-boom) and 1980
(extension of education after high school). Inequalities concerning sc
hooling are stable at their lowest level in recent years. The educatio
nal investment fluctuates according to demographic reasons but its tre
nd is given by the increase of women school attendance, the increase o
f their participation in the labour force and the general lengthening
of higher education after the second world war.