Types of FDI and determinants of affiliate size : the classification makes the difference

This paper deals with the measurement of motives for foreign direct investment (FDI). Due to a lack of information, several indirect measures exist in order to classify multinational firms into the two main types of FDI. While vertical foreign direct investment (VFDI) refers to the international fragmentation of the production process for cost-saving reasons, horizontal foreign direct investment (HFDI) is performed in order to gain access to new markets. One common approach to identify the dominant reason for firms to go abroad is to compare the industry affiliation of the investing company in the home country and the subsidiary in the target country. The question arises as to how reliable this measure is for identifying FDI motives. The IAB-ReLOC survey allows a profound investigation on the issue of classifying the motives of the firms for going abroad into vertical and horizontal FDI. Apart from industry affiliation data applied in conventional approaches to categorize FDI types, the survey data also includes a self-assessment of the firms with respect to the main motive for investing in the neighboring country, and information on intra-firm trade concerning the flow of intermediate inputs between the German headquarters and the Czech affiliates. Against the background of featuring a well-grounded database, we shed a light on the relevance of productivity in the German-Czech FDI relations. We pursue a reference group approach by comparing German multinational firms that have an affiliate in the Czech Republic to German companies without direct investment abroad. The data provided by the German multinationals enables us to investigate the size of FDI under the aspect of the number of employees in their Czech affiliates. By applying a two-step Heckman procedure, we control for sample selection bias: in the first stage we analyze the extensive margin of FDI, i.e. the probability to select into the group of multinational investors. The second stage examines the relationship between productivity and the intensive margin of FDI. We find evidence that productivity is not only a crucial factor for the decision to invest in the neighboring country, but plays also a relevant role for the number of employees in the Czech subsidiary. Differences are revealed between direct and indirect measures of FDI types. The size of horizontal investments is significantly affected by productivity only in the case of classifications that are based on survey responses. This result confirms theoretical expectations and previous empirical literature by standing in marked contrast to the outcome for indirect measurement concepts. Our finding leads us to the conclusion that one should be more cautious in interpreting differences between vertical and horizontal FDI when using approximative classification concepts.