The Linder hypothesis for foreign direct investment revisited