tributed from 35 (Dominica) to 100 (St. Kitts and Nevis), with the not- able exception of Micronesia (ranked 123rd). The slight tendency of smaller countries to have weaker health sys - tem performance emerging from Table 4.9 seems to be due to small countries above our threshold of 500,000 inhabitants and not due to the relatively strong performing VSC in Table 4.10. The average rank of the 20 countries (with the exception of Liechtenstein) – 55.6 – corroborates this notion, although lower income VSC are disregarded. VSC clearly do not have worse performing health systems in comparison to larger coun- tries when we control for per capita income. But does this astonishing performance mean higher costs for VSC? The central question is whether there is a negative relationship be - tween health expenditure and country size; or stated differently: Do smaller countries have higher expenditure on health? Recall Hypothesis 2, where we formulated the theoretical expectation. Given the fact that there is no influence of country size on health system performance, we should have higher costs and, therefore, higher health expenditure for smaller countries due to economies of scale effects, according to Hypothesis 2. There are various measures for health expenditure which are of interest in our context, like total health expenditure in % of GDP, public health expenditure in % of total health expenditure or public health expenditure in % of total public expenditure. Table 4.11 displays the results for the according regressions. Table 4.11 reveals that there is no statistically significant relation - ship between country size and health expenditure as a percentage of GDP, even if we control for per capita GDP. Therefore, size does not de- termine health expenditure in any direction, and smaller countries do not have higher health expenditure, which is a finding that clearly con - tradicts our expectations. Size, furthermore, does not play a role in ex- plaining the fraction of total public expenditure spent on health. Hence, two important measures of health expenditure are not related to country size.We only find a significant and unambiguous influence of country size on a structural variable of health expenditure. The fraction of public expenditure on health of the total health expenditure decreases with country size. Larger countries seem to prefer to rely on private financ - ing at least of parts of the health system. In smaller countries, most of the overall health expenditure is financed publicly. 114 
Very small countries: organizational choice and international outsourcing