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

-29-

Chart 4 - Impulse response analysis in SFAVAR (dummy approach)

In the next step, we examine the stability of the global structural shocks. Under
the null hypothesis that there is no break in the variance, E | η_t(τ)| is constant. We
therefore test for a break by implementing the Quandt-Andrews test in the regression of
| η_t (τ) | against a constant, using homoskedastic standard errors (which are valid under
the null). Accordingly, apart from the short-term interest rate innovations, there is no
indication of sharp structural breaks. The estimated break date for global interest rate
shocks is Q1 1991. Given that such a sharp break may still be part of an ongoing trend,
we additionally test for more gradual changes in the global short-term interest rate
shocks. The regression is augmented with a time trend (Stock, Watson 2002). The null
hypothesis of no sharp break is not rejected again. The coefficient sign of time trend is
significantly negative, thereby reflecting that the magnitude of the shocks is decreasing
over time.

<<   <   29   30   31   32   33   34   35   >   >>