Two of the main puzzles in international economics are the consumption and the portfolio home biases. They are empirically related: countries that are more open to trade also have more internationally diversified portfolios. In a two-country stochastic equilibrium model, I prove that introducing trade costs in goods market alone, as suggested by Obstfeld and Rogoff 2000, is not sufficient to explain these two puzzles simultaneously. On the contrary, for reasonable parameter values, trade costs create a foreign bias in portfolios. To reconcile facts and theory, I introduce a combination of small frictions in financial markets and trade costs in goods market. The interaction between the two types of frictions determines optimal portfolio allocation. When trade costs increase, competition in the goods market softens and the volatility of domestic income falls. Facing lower risk, investors have less incentive to pay the financial transaction cost and increase their holdings of domestic assets. The model correctly predicts that the larger the home bias in consumption, the larger the home bias in portfolios.