Gunnar Bårdsen and Ragnar Nymoen 911R∗ECU interest rate. For the period 1967(1)–1986(3): effective interest rate on
foreign bonds, NOK basket weighted. For the period 1986(4)–1996(4): ECU
weighted effective rate on foreign bonds.
RLAverage interest rate on bank loans.
RBYield on six year government bond, quarterly average.
R∗BYield on long-term foreign bonds. NOK basket weighted.
PConsumer price index (CPI).
PC“Core” CPI.
P∗Consumer prices abroad in foreign currency.
PECPI, electricity, fuel and lubricants.
PO$US oil price, per barrel Brent–Blend.
PIPrice deflator of total imports.
PI∗Producer price index, trading partners.
YTotal value added at market prices in the mainland economy (defined as total
economy minus North Sea oil and gas production and international shipping).
Fixed base year (1991) prices. Mill. NOK.
ZMainland economy value added per man hour at factor costs.
T1 Payroll tax rate, mainland economy.
URegistered rate of unemployment.
WNominal hourly wage costs in the mainland economy. NOK.
In addition, there is a step dummy, accounting for the introduction of inflation
targeting:
IT=0 until 2001( 1 ), 1 from 2001( 2 ).Notation for estimation and misspecification tests
In Table 17.1, the estimation method, which is either OLS or full information max-
imum likelihood (FIML), is indicated in the first line below each equation, along
with the sample size (number of quarterly observations), which is denoted byT,
and the residual standard error (σˆ). For equations estimated with OLS, statistics
for residual autocorrelation and ARCH are reported in the second line below the
equation. As indicated by the notation, these two statistics areF-distributed under
their respective null hypotheses. For example,FAR( 1 − 4 )(4, 44)in the exchange rate
equation denotes theF-distributed test statistics with 4 and 44 degrees of freedom
for the null hypothesis of no autocorrelation against the alternative of fourth-
order autocorrelation. In the third line below the estimated OLS equations, we
report the chi-square-distributed test of residual normality, and the F-distributed
test of heteroscedasticity due to squares of the regressors. For the equations esti-
mated with FIML, systems versions of the misspecification tests are reported and
are indicated by the extra subscriptvec,asFvec,AR( 1 − 4 ). The numbers in brack-
ets arep-values for the respective null hypotheses. These, as well as the other
standard diagnostics tests, are explained in Doornik and Hendry (2007a) (single-
equation diagnostics), and Doornik and Hendry (2007b) (system and simultaneous
equations diagnostics).