The results of Table 5.4 are not easy to interpret at first sight. Ana - logous to our discussion on the whole set of VSC, we obtain a highly di- versified picture when we only consider the high-income VSC. This sec- tion is designed to highlight some common patterns and to reveal a few astonishing results which contradict conventional wisdom or theoretical expectations and prejudices concerning VSC. Contrary to the widely held belief, we have VSC with a high living standard which have a large and profitable industrial sector – despite the small home market and the associated diseconomies of scale –, we have VSC which are clearly not dependent on financial services or off-shore banking, and we have quite some high-income VSC with a rather fractionalized, inhomogeneous po- pulation with regard to ethnicity and languages spoken – despite the common understanding that homogeneity is one of the major advanta- ges of 
VSC. 5.2.2.9.1 History and socio-economic factors Starting with the assessment of the duration of independence, it is note- worthy that five of our eight VSC in Table 5.4 have been independent at least since 1815, which is a remarkably long time given the small politi- cal and military potential of VSC and the frequent map-changing wars in Europe. Comparing the time of independence with those of low-income VSC, a significant difference emerges. High-income VSC exhibit on average a longer period of independence than low-income VSC (Mann- Whitney-U-Test; two-sided, p = 0.005, N=21).113 We do not believe that traditional VSC are more capable of exploit- ing the benefits of sovereignty per se due to longer experience. It is there fore more likely that a concomitant variable exists which influences both welfare levels and the date of independence. The latter argument seems to offer a good guess, given the fact that many VSC which achiev - ed the status of independence since 1960 were former colonies. The 153 
Sources of welfare in VSC 113Note further that a univariate regression with per capita GDP as dependent and the duration of independence as independent variables yields a significant result on the 5% significance level. We did not enclose this variable in the multiple regressions above because of the lack of a theoretical rationale. Besides, there are some difficul- ties when verifying exact dates of independence, for instance, due to rules of country succession. We would however conjecture that the effect is much less pronounced, if significant at all, when all countries and not only VSC are under consideration.