iv. In the “with Structural Funds” simulation, we set the externality elasticities to
a standard set of values for all four models. These are in the mid-range found
in the international literature, and both the output and factor productivity
elasticities are set at 0.20 (i.e., a one per cent rise in the stock of physical
infrastructure or of human capital will increase the level of output and the
level of factor productivity in the medium term by 0.2 per cent). We relax this
assumption later when we carry out a sensitivity analysis.
v. We then “extract” the Structural Funds 94-99 public policy shocks (i.e., EU
and domestic public expenditures) from the above simulation, i.e., we set the
Structural Funds 94-99 expenditures at zero and re-simulate the model. No
other changes are made., and no attempt is made to design a “substitute”
domestically funded public investment programme that would have replaced a
“missing” Structural Funds 94-99. This is a very artificial assumption, since
in the absence of Structural Funds 94-99 there almost certainly would have
been substitute domestically funded public investment programme, albeit
smaller in magnitude.
vi. Ideally we should use the actual ex-post realised Structural Funds
expenditures. But these were not available for every country or region,
disaggregated by priority and on an annual basis.9 In the interests of
uniformity, we have used the planned Structural Funds expenditure data as
contained in the Structural Funds 94-99 treaty documents. While these give a
fairly accurate total for the expenditures, they do not always give an accurate
picture of the ex-post scheduling of the expenditures. However, this is only an
important issue in the case of Greece, where the planned even spread of
expenditures over the six years 1994-99 was actually implemented in a revised
fashion that involved “back loading” by re-programming. However, the totals
are similar to the planned expenditures. Consequently, the medium-term
macro impacts will be the same as for the “actual” expenditure pattern.
9 Complete ex-post CSF 1994-99 data were only available for Northern Ireland , Portugal and Ireland.
14