The Importance of Global Shocks for National Policymakers: Rising Challenges for Central Banks

-31-

Table 6 - Andrews-Quandt breakpoint tests for factor loadings

Note 1: Heteroskedasticity-robust version of Maximum LR test

Note 2: *** Indicates significance at 1% level, ** at 5% level, * at 10% level

However, as already noted, the Andrews-Quandt test focuses on rather simple
one-off breaks. Such an assumption is not justified, if gradual changes in coefficients and
residuals occur. In addition, there can be temporary outbursts of volatility. Tests for a
single change in volatility are then misleading, as the one-off break will be dated either at
the beginning or at the end of the extremely volatile time period. We therefore estimate
rolling regressions for factor loadings, concentrating on GDP, money and house prices.
Given that our sample covers a comparatively short time span of 24 years, we opted for a
rolling 12-year window.

Selected results can be found in the following charts in which the solid line equals
the factor loading and the two dashed lines are the 95 percent intervals obtained by a
simple residual bootstrap with 500 draws (see also Tables A3 to A5 in the Appendix). For
example, the figures for Q1 2000 are the results based on the rolling regression from Q1
1988 to Q4 1999. A clear-cut pattern across countries and variables does not emerge. In
some cases, the factor loadings increase over time, in some not or even decline.
Concerning global liquidity, the factor loading for EMU M3 more than doubles, whereas
the impact on US M2 also rises but less markedly. The same diverse pattern also holds
for the relation between national and global GDP and national and global house prices,
respectively.

Interestingly enough, US variables in general seem to be influenced increasingly
by common forces. For example, the US factor loadings both for global GDP and house
prices have risen significantly. However, at this point, it is not clear whether the US has
really become more prone to influences from abroad. We will deal with this issue in the
next chapter by conducting several robustness checks.¹¹ In contrast, the effect from

¹¹ Given the fact that our common forces for money, GDP and house prices are estimated on the basis of
only five variables, the reverse causality may hold, i.e. development in the US has recently become more
important for the global economy.

<<   <   31   32   33   34   35   36   37   >   >>