Purpose – The aim of the study is to investigate the relative significance of the determinants of disaggregated capital flows (foreign direct investment and portfolio flows) to five developing countries, across different time horizons. Design/methodology/approach – An empirically tractable structural VAR model of the determinants of capital flows is developed, and variance decomposition and impulse response analyses are used to investigate the temporal dynamic effects of shocks to push and pull factors on foreign direct investment and portfolio flows. Findings – Estimation of the model using quarterly data for the period 1976-2001 provides evidence supporting the hypothesis that shocks to real variables of economic activity such as foreign output and domestic productivity are the most important forces explaining the variations in capital flows to developing countries. Research limitations/implications – These findings highlight the concomitant need for policy makers in developing countries to design domestic policy that accounts for both external and internal shocks to real variables of economic activity. Originality/value – Previous empirical studies on the determinants of capital flows to developing countries have mostly examined the capital flow variable in aggregate, and have largely overlooked the possibility that the relative significance of estimated coefficients of such determinants may vary across time horizons.