This paper studies the effects of government spending under limited international capital mobility, as featured by most developing countries. While external financing of government debt mitigates the crowding-out effect, it generates real appreciation, which contracts traded output and lowers the fiscal multiplier in the short run. The decline of the multiplier is larger when facing debt-elastic country risk premia. Also, government spending is more expansionary with more home bias in government purchases, more sectoral rigidities, and a less flexible exchange rate. Whether the twin-deficit hypothesis holds depends crucially on the extent to which government deficits are financed externally.