In this paper we examine the relationships between two sets of three variables: Swedish real exports, Swedish real GDP, and foreign real GDP in one set; and Swedish real exports, Swedish total factor productivity, and foreign real GDP in the other set. The foreign real GDP facing Sweden is proxied by total OECD real GDP minus Sweden's real GDP. Multivariate tests for integration and cointegration show that the variables in each model are cointegrated. We also perform Granger causality tests on these variables in our examination using the Toda-Yamamoto procedure. We discover bi-directional causality between Swedish real exports and Swedish real GDP (or Swedish total factor productivity). Foreign real GDP is shown to Granger cause Swedish real exports, but no significant causation of foreign real GDP on either domestic GDP or total factor productivity was found. A change in foreign real GDP thus appears to affect Swedish output and productivity only indirectly, through changes in Swedish exports.