between economies are apparent. Thus, the largest impact as measured by the
cumulative multiplier appears to have been in East Germany, which might be
somewhat surprising. However, this effect appears to have been dominated by other
negative effects yielding the modest overall economic performance of East Germany.
The lowest impact was found for Greece, and this might be linked to the relatively
low level of economic openness of that country.
Finally, the direct benefits arising from the Structural Funds are only part of a much
wider picture. The real long-term benefits of the Structural Funds are shown to be
associated with the way in which each economy (region) responds to opportunities
arising in the rest of the country and the EU as a result of the Single Market rather
than from the Structural Funds in isolation. This emphasises the need to work within
the wider “global theory” of macro modelling rather than the narrower “theory of
action” that tends to motivate policy makers who are focused on the role of specific
Structural Funds programmes.
Of course the HERMIN framework is not without weaknesses. For example the
evaluations are dependent on the chosen externality elasticities. Increasing the size of
the externality elasticities boosts the impact of the Structural Funds programmes.
Since precise values of the elasticities are not known, a range of possible Structural
Funds impacts must be considered. However, sensitivity analysis suggests that the
results are relatively robust.
Another possible weakness is the relatively high level of sectoral aggregation. For
policy makers, particularly those interested in industrial policy a breakdown of for
example manufacturing into sub-sectors would be interesting as certain sectors may
benefit more or less form the Structural Funds. Against that one has to consider the
issue of data availability and analytical complexity that further disaggregation would
introduce. Nevertheless, efforts are underway to disaggregate the sectors further.
The various models are currently not explicitly linked to each other. The exceptions
here are migration and commuting flows which in the case of the regional models are
incorporated and the exogenous demand linkages that are also a feature of the model.
However, particularly at the regional level a more explicit linkage between the models
25