New residential construction is significantly more procyclical in emerging markets than in developed countries, although the correlation between aggregate investment and output is similar across emerging and developed countries. This paper shows that a multi-sector stochastic growth model with a housing production sector can explain this fact. The key feature of the model is that housing demand depends on the cyclical behavior of consumption of tradable goods, which is much more volatile in emerging markets. Therefore, when a positive productivity shock hits the economy, the larger response of consumption of tradable goods implies that it is more attractive for consumers in emerging markets to purchase housing than it is for consumers in developed countries. This paper considers various factors that contribute to the large variability of consumption in emerging markets, and finds that larger trend growth rate shocks in emerging markets than in developed countries are quantitatively important. The reason is that a positive productivity shock signals even higher productivity in the future with large growth rate shocks, so the current consumption response is large and the return to housing investment is high. While qualitatively the model matches the differences in the cyclicality of new residential construction across emerging markets and developed countries, quantitatively the model underestimates this comovement and the volatilities in housing investment in emerging markets.

Furthermore, international interest rate shocks highly correlated with productivity shocks are very important in explaining the large swings in housing investment in emerging markets. Interest rate shocks work through three channels to affect housing investment: the direct mortgage rate' effect, the indirect effect through increasing non-housing consumption and the supply effect due to the working capital constraint. Quantitatively, the direct mortgage rate' effect is the most important channel. When the housing asset acts as collateral to reduce household's financing costs, it provides an empirically important mechanism to amplify and propagate interest rate shocks over the business cycle. The reason is that housing prices and interest rates reinforce with each other to generate more procyclical housing investment and more volatile consumption and output.