This paper argues that residential investment is much more volatile
than GDP because the demographic variable that determines specifically
housing demand (growth of population aged 20-49) is itself
highly volatile. First, we depart from the literature on housing cycles
which has mainly focused on explaining the cycle of housing prices
by changes in financial conditions. Instead, we study the cycle of real
residential investment as a share of real GDP in 20 OECD countries
since 1980, and show that it is closely associated with the growth dynamics
of population aged 20-49. Second, the paper develops a new
method to uncover the causal effect of the growth of population aged
20-49 on housing construction. We use past demographic data as an
instrument to avoid the potential endogeneity bias between migrations
and the housing cycle. The instrument is strong in countries where net
migration is low. Overall, we find that a 1% increase in the population
aged 20-49 increases the ratio of residential investment to GDP
by 1.3 pp. Demographic changes appear to be a better predictor of the
cycle of the residential investment rate than any other macroeconomic
variable we control for. (JEL E32, J11, R21)