The last few years have seen important advances in methodologies for testing for both
short-run and long-run impacts of fiscal policy on growth but with results from short and
long-run approaches generally difficult to compare. In addition, the ‘long-run’ in
most studies is typically identified using cross-section or panel methods applied to
period-averaged data, with assumed common impacts across countries. As a result
long-run fiscal impacts are identified rather crudely and estimated, or assumed, to take
5–10 years or more to affect long-run output growth rates. However, as Pesaran and
Smith (1995) and others have argued, assuming incorrectly that such parameters are
homogeneous across countries is likely to bias results.