| dc.description.abstract | World trade today is fragmented, and compared to conventional statistics trade in value added provides a better measure on the value created in a country from trade. This because conventional statistics are subject to a double-counting when imported inputs are used in production of exports and therefore inaccurate as measure of the value from trade. This thesis investigates the relationship between trade in value added and quality of international trade statistics. To calculate TVA an international input-output table is necessary, and this table is created from national input-output tables, combined with bilateral data on international trade. The challenge occurs when we find large deviations in data on international trade. Intuitively, reported imports in Norway from the Netherlands should equal reported exports from the Netherlands to Norway, after adjusting for transport costs. Instead, we find large deviations in such mirror data , between two values reported on the same flow. Our analysis reveals that average import value reported in Norway exceeds the export value by 100%, and for Asian countries is it often even higher! Such deviations in mirror data affect trade in value added in two ways; first, we need to harmonize mirror values to create an international input-output table and calculate value added. Second, extra costs from services are reflected in some of the deviations, and this should be accounted for as part of the value chain. We review theoretical and empirical results in the literature on trade in value added, followed by an empirical study on Norwegian imports from 114 countries, and a comparative analysis on the Netherlands trade with 182 countries. The goal of this analysis is to reveal systematic explanations in mirror data discrepancies, due to a country s position in trade. Norway and the Netherlands make good candidates; Norway being small and periphery, and the Netherlands having a central position with a large port in trade. To explain discrepancies in mirror data, we use a method of decomposing them into price and quantity effects, finding systematic differences between the two countries. For Norway price effect increases with distance to partner country, but quantity effect is prominent in explaining discrepancies. The Netherlands has many observations of exports reported as higher than matched imports, likely explained by its position as an entrepôt country .Letting such systematic differences influence harmonization of data can improve calculations of trade in value added. | eng |