Prepared by Luis Catão.
Ideally, it would also be useful to have estimates based on the production function; since the latter incorporates information on the growth of the capital stock and the labor force, it normally provides a further consistency check on the other two methods. In the case of Ireland, however, the lack of a comprehensive series on capital stock and the difficulties in projecting the growth of the labor force (due to the volatility of immigration and sharp changes in participation rates) prevent a meaningful use of the production function approach for the present purposes.
See Chapter 5.
Defined as GDP minus agriculture minus manufacturing output.
There are two other reasons for considering this alternative indicator. First, increases in productive capacity in those services tend to respond to demand pressures in a much slower fashion than in other sectors. Second, service data are much less subject to possible distortions related to transfer pricing.
The OLS regression is run from 1979, with a trend break in 1987. The choice of 1979 and 1987 as turning points is clear in the Irish case. The first marks the end of monetary union with the United Kingdom, Ireland’s ERM membership and the second oil shock. The period starting in 1987 is characterized by debt consolidation, greater macroeconomic discipline and rapid economic growth.
Consumer price inflation during those years was slightly above 3 per-cent.
The growth of the non-tradable index for 1995 was projected according to the elasticity of non-tradable output to GNP during 1979–94. Non-tradable output for 1995–97 was then extrapolated using the staff projections of real GNP for those years.
The HP filter is based on the minimization of the following loss-function associated with deviations of log of a variable (y) from its trend (y*):
where the parameter λ determines the degree of smoothness of the filter and is usually set to 1,600.