actually sourced out (given a certain degree of preference homogeneity between the VSC and adjacent countries if necessary). One could also imagine testing whether goods with a very limited level of economies of scale are produced in-house in VSC, but this is not the question we are interested in, because we aim to explain how VSC cope with their size disadvantage and the according higher costs of public good production. Note that a comprehensive assessment of all the goods listed in Table 4.3 is not possible due to serious operationalization problems for some of them and problems with assembling the necessary data. We therefore choose at least one publicly provided good from each category in Table 
4.2. 4.3 Organizational choice: theoretical expectations versus reality In this section Hypothesis 1 is tested for a set of public goods in order to learn more about the possibilities of international outsourcing in VSC. We present one table for each of the selected public goods. Due to constraints on data availability, data harmonization and operationaliza- tion problems, it is not possible to test for all the public goods mentio- ned in Table 4.3. Nevertheless, the analysis is quite comprehensive and an overall picture of public good provision and organizational choice in VSC emerges. The empirical analysis in this sections builds on public goods in 21 VSC which fulfill the following criteria: –The number of inhabitants is smaller than 500,000. –Per capita GDP or GNP is higher than $ US 2,000. –We consider only fully sovereign and internationally recognized countries. The criteria employed are rather restrictive in order to avoid arriving at unclear results, which may be the consequences of too heterogeneous a set of countries. Note that nevertheless VSC in five continents are rep - resented in our set. Our country set comprises Caribbean islands and Oceanic archipelagos as well as landlocked European VSC and the only VSC – according to the above-mentioned definition – in Asia and Africa, Brunei and the Seychelles, 
respectively.89 
Organizational choice: theoretical expectations versus reality