-25-
the technology innovations do not have an impact on the global structural shocks within
the first four quarters, i.e. c0 , c1 , c2 , c3 and c4 are jointly equal to zero. A constant in
equation (8) is neglected since its influence proved to be non-significant. The results
indicate that the null hypothesis is rejected for five of the seven global innovations. Apart
from the global money and the short-term interest rate shock, all innovations in the
FAVAR are contaminated by technology shocks. This is true independent of whether
total hours worked are modeled as stationary time series or integrated of order one.
Interestingly, the coefficient signs are not always in line with economic theory. For
example, global technology shocks have a positive impact on global inflation. It is not
clear whether this result is triggered by another common force behind technological
innovations or due to the specification suggested by GaH (1999).
Table 2 - Testing for omitted global technology shocks in the SFAVAR
Global structural shocks |
Hours (I(1)) |
Significant Coefficients |
Hours (I(0)) |
Significant Coefficients |
GDP____________ |
3.04** |
c0=-0.15**; c1=0.14**; c2=0.12* |
3.29*** |
c0=-0.16**; c1=0.14**; c2=0.13* |
Inflation |
2.04* |
c1=0.61** |
2.25* |
c0=0.54**; c1=0.64* |
Commodity prices |
1.95* |
c1=0.34* |
2.22* |
c0=0.40**; c1=0.35* |
House prices |
2.39** |
c1=-0.15**; c3=0.11* |
196* |
c1=-0.11* |
Monetary liquidity_______ |
182 |
— |
137 |
— |
3M interest rates_______ |
1.16 |
— |
1.26 |
— |
Share prices |
2.96** |
c0=-0.28*; c1=-0.40**; c2=0.40** |
2.83** |
c0=-0.30*; c1=-0.35*; c2=0.41** |
Note 1: F-statistics from Wald tests
Note 2: *** Indicates significance at 1% level, ** at 5% level, * at 10% level
The same type of Wald tests in equation (8) are repeated for long-term interest
rate shocks. Hence, a principal component analysis for 10Y government bond yields is
done.7 Accordingly, PC1 amounts to a high 66% (PC2: 13.9%). Again, we “re-construct”
the PC1 in levels by setting the global interest rate factor zero in Q1 1984 and calculating
the cumulative sum. An equation is estimated with the global interest rate factor as
dependent variable and the common forces used in the SFAVAR (GDP, inflation,
commodity, house prices, liquidity, short-term interest rates and share prices) as
independent ones. As is the case in our baseline FAVAR, a lag length of two is chosen.
The obtained residuals are taken as a proxy for the global interest rate disturbance for the
Wald tests. Accordingly, common long-term interest rate shocks seem to play an
important role in the global economy. Of the seven common disturbances in the
7 In the case of Japan, the average maturity for government bonds is used for data availability reasons.