THE ECONOMETRICS OF MACROECONOMIC MODELLING

Mainland GDP output yt

The model for Ayt is adapted from the ‘AD’ equation in Bardsen and Klovland (2000). The growth in output Ayt is in the short run a function of public demand Agt, and growth in private demand—represented by growth in real private credit Acrt. Moreover, there is an effect from the change in the real exchange rate in the period after the deregulation of currency controls in Norway in 1990(2).

Ayt = 1.16 — 0.39Ayt_ і + 0.29Agt + 0.49Acrt_ 1 (0.30) (0.07) (0.06) (0.12)

— 0.17 ecmy t + 0.41 (s ■ A(v + pw — p))t-2 + 0.07Ydumt (0.05) ’ (0.12) (0.01)

— 0.06 Seasonal^ 1 — 0.07 Seasonal^2 — 0.03 Seasonal^3 (0.003) (0.005) (0.004)

(9.8)

T = 1972(4)-2001(1) = 114

d = 1.21%

Far(i-5)(5, 99) = 0.84[0.53]

X2 normality (2) = 0.78[0.67]

FHETx2 (14, 89) = 0.48[0.94].

(Reference: see Table 9.2. The numbers in [..] are p-values.)

The equilibrium-correction mechanism of aggregate demand, denoted ecmy, t is defined as:

ecmy, t = yt-і — 0.5yw—i — 0.5gt_i + 0.3(RL — 4Ap)t-i,

where the long-run steady-state is determined by real public consumption expenditure (gt), real foreign demand, which is proxied by the weighted GDP for trading partners (ywt), and the real interest rate on bank loans rate (RL — 4Ap)t, where RLt is the nominal bank loan rate. The estimated adjust­ment coefficient of —0.17, suggests a moderate reaction to shocks to demand. The estimated equation also includes a constant and three seasonal dummies and in addition the dummy Ydumt = [i75q2]t is required to whiten the residuals.

9.3.1 Unemployment ut

The dynamics of unemployment Aut display strong hysteresis effects, with very sluggish own dynamics. Also aggregate demand shocks A4yt and changes in the real wage A(w — p)t have significant short-run effects. Moreover, there are significant effects of change in foreign demand Ayw, and the share of the

workforce between 16 and 49 years old N16-49,t.

Aut = — 1.23 + 0.34Aut_ і — 0.06 ut_ і — 1.83Д4У (0.49) (0.07) (0.02) (0.29)

+ 1.30Д(ад — p)t-i —2.63Дут—2 +1.78Ni6-49,t + 0.22Udumt (0.52) (1.03) (0.71) (0.04)

+ 0.41Seasonalt_ 1 + 0.10Seasonalt_ 2 + 0.29Seasonalt_ 3 (0.03) (0.02) (0.02)

— 6.46chSeasonalt_ 1 — 7.55 chSeasonalt_ 2 — 4.34 chSeasonalt_ 3. (0.49) (0.34) (0.39)

(9.9)

T = 1972(4)-2001(1) = 114

a = 5.97%

Far(i-5)(5, 95) = 0.69[0.63]

Abnormality (2) = 1.91[0.38]

FHETx2(23, 76) = 2.21[0.005].

(Reference: see Table 9.2. The numbers in [..] are p-values.)

There are two sets of seasonals in this equation. chSeasonalt is designed to capture a gradual change in seasonal pattern over the period:

Moreover, a composite dummy variable Udumt = [i75q1 + i75q2 — i87q2]t is required to whiten the residuals.

Summing up, the unemployment equation in essence captures Okun’s law. An asymptotically stable solution of the model would imply U = const + f (Ду), so there is a one-to-one relationship linking the equilibria for output growth and unemployment.