Empirical essays on Regional Specialization and Trade in Sweden

Abstract: This thesis deals with the empirical testing of trade theories, traditionally used to analyze international trade, using regional data on production and trade. This thesis consists of three main essays. In the first essay, we analyze Swedish regional production. In the first part, we focus on the effects of comparative advantage in predicting regional manufacturing production. In so doing, we use a combined Heckscher-Ohlin (HO)-Ricardo approach and analyze the effect of technological differences and factor endowments. The result of our empirical test offers evidence supporting our hypotheses, namely that relative endowments of human and physical capital as well as technical differences have significant effects on Swedish regional production. In the second part of this essay, we combine the HO framework with economic geography to empirically analyze the effects of factor endowments and market size on Swedish regional manufacturing. The result from this analysis shows support for our hypotheses, namely that factor endowments as well as market size have significant effects on the pattern of manufacturing production. In the second essay, we perform a test of the Hecksher-Ohlin-Vanek (HOV) model on Swedish regions’ trade. In so doing, we incorporate input-output analysis first introduced by Leontief to estimate the factor content of each regions’ trade. This is then compared to the theoretical predictions. While previous research has offered little empirical support for the standard HOV model, recent studies have shown that small modifications in the underlying model lead to improvements in the empirical fit. Following recent research, we modify the standard focus of the model to focus on the factor content of Swedish region’s trade. In the first part of the analysis, we use a traditional approach, in using one technology matrix to estimate the factor content of all regions’ trade. Surprisingly, the result of this analysis shows no support for the model in predicting the factor content of trade. In the second part, we introduce region-specific technology matrices, which are used as basis for our estimations. We argue that, due to the construction of input-output matrices, assuming these to be identical across regions does not only imply identical technologies. This in effect implies that each country or region has identical production and import of goods used as input. Although, the result of this analysis shows an increase in the overall fit of the model, the empirical support is still weak In the third essay, we analyze the geographical concentration of industrial activities in Sweden. First, we calculate locational Gini-coefficients as introduced by Krugman (1991) to study the relative concentration of Swedish industries. The result of these calculations shows an overall concentration of industries, however not more so for skill-intensive industries than others. Secondly, we empirically analyze the effect of factors stemming from new trade theory and economic geography frameworks on the concentration of industrial activities in Sweden. Contrary to the theoretical predictions, the result of our empirical analysis reveals no support for skill-intensity, use of intermediate goods or economies of scale in production having significant effects on the concentration of manufacturing industries. However, the use of intermediate goods used in production was shown to have a positive and significant effect on concentration when service industries were included in the sample.