account household savings rates across regions. An outstanding feature is the relatively low US
household savings rate. To better capture this phenomenon we set the US rate of time preference to
1.5%.
Production: The output elasticity of labour is set equal to an average wage share across OECD
countries. An annual depreciation rate for capital of 6% is assumed. The adjustment cost parameter
is difficult to pin down on the basis of first moments of the data. It has been shown in the business
cycle literature, however, that this parameter is crucial for determining the relative volatility of
investment (see, for example, Mendoza (1991)). It is therefore set in such a way as to make
investment about 3 times as volatile as GDP.
Labour market: Wage setting is characterised by three parameters, the level of the
unemployment replacement rate, the parameter χ which determines “bargaining strength” of
workers and the elasticity of wages w. r. t. unemployment. Since our starting point is 1970 where
we observe low unemployment rates across the world we assume identical parameters for wage
setting for all regions. Unemployment benefits are set such that they amount to roughly 30% of
gross wages, which is around the order of magnitude found for a weighted average of EU countries
and the US. The elasticity of wages with respect to unemployment is set to .5 which again
corresponds to a weighted average of elasticity estimates found for the EU and the US. The
bargaining strength is set to .5. The indexation of unemployment benefits to gross wages is chosen
such that an increase in labour taxes or social security contributions by 1% leads to an increase of
the unemployment rate of around .3% in the long run. This is about the average value reported by
various empirical studies with values ranging from practically zero (Blanchard and Wolfers
(2001)) to a value of .5 obtained by Daveri and Tabellini (2000) for example.
International financial markets: Little is known about the degree of international capital
mobility. Earlier simulation exercises have, however, shown that full capital mobility between
EU15, US and Japan on the one hand and the ROW would have led to net foreign asset positions
which would by far exceed the observed international imbalances. It is assumed that international
financial markets are incomplete by introducing a risk premium which depends on the net foreign
asset position of the respective region. We impose a small risk premium. Previous research
(Roeger 2003) has shown that higher risk premia for the non OECD regions are necessary such
that the model approximately fits the observed net foreign asset position of the fast and slow
ageing RoW. For example, without risk premia, the fast and slow ageing RoW economies would
exhibit net foreign debt ratios in the order of magnitude of about 120%, which is nearly ten times
larger than the observed level in 2000.
3. Standard Policy Experiments
The pension reform scenarios analysed below will be combinations of changes in government debt
combined with changes in social security contributions (combined with changes in the generosity
of the pension system). In order to better understand the total effect it is useful to look at the
individual effects separately by conducting two standard simulation experiments. This also allows
us to see how the effects in this model relate to results in the literature. We analyse two
experiments, a permanent increase in government debt and a reduction in pensions.
80