exists. Only trade openness enters individually significantly.
The findings for the average stock of foreign capital liabilities and its sub-components can
be summarised as follows. Considering the infrastructure variables separately, roads do not
seem to contribute to the attraction of new international capital. By contrast air-departures
and international telephone circuits have a positive and statistically significant relationship with
the average total liability stock for the period 1990-95. Air-departures also show a positive and
statistically significant influence on FDI, while international telephone circuits have a positive
impact on the attraction of new debt. Overall, the joint impact of the chosen infrastructure
variables on the stock of capital inflows is not statistically significant. An exception is provided
by the sub-component on FDI. Here a statistical significant impact of infrastructure on new FDI
is found.
3.3.2 1990-95 Cross-Section Analysis of the Flow Data
The analysis of the average capital flows illuminates interesting effects on FDI flows. All infras-
tructure variables enter with a positive sign into the bivariate specification of FDI flows in Table
8. However, as seen above for the stock of FDI, only air-departures are statistically significant at
the one percent level (column (1)). A 2.5 percentage point increase in the level of air-departures
raises the average inflow of FDI liabilities by 0.1 percentage point. Note that 33 percent of the
cross-country variations in FDI flows are explained by this variable. The result remains valid
when country differences, depicted in columns (4) to (6), are controlled for. However, the inclu-
sion of the remaining regressors in columns (7) to (9) wipes out the individual significance of the
air-departure variable and leaves the other infrastructure variables unchanged. In all three spec-
ifications of columns (7) to (9) trade openness has a negative and statistically significant impact
on the specifications. The same is true for natural resources in columns (8) and (9). In these
columns the dummy variable concerning being landlocked also enters individually significantly
and with the expected negative sign. The results are similar to the ones obtained for the stock of
FDI. Interestingly, looking at the joint influence of infrastructure in column (10), an individually
significant relationship of air-departures and international telephone circuits is found. While the
former reveals a positive sign, the latter is negative. A test for the joint significance of the three
infrastructure variables has a p-value of 0.001. The variables are jointly highly significant. The
individual impact of openness and natural resources remains stable. Overall, for the flow data
joint significance of the infrastructure variables is found for FDI flows. An individually positive
relationship exists for air-departures and FDI inflows, as documented above.
11