Palgrave Handbook of Econometrics: Applied Econometrics

904 Macroeconometric Modeling for Policy

2003

–2

0

2

4

6

2004 2005 2006 2007

2003

–2.5

–5.0

0.0

2.5

5.0

7.5

10.0

2

1

3

4

5

6

7

–4

–8

–20

–10

0

10

20

0

4

8

12

4

0

8

12

16 30

95

90

85

80

100

105

110

2004 2005 2006 2007

2003

1

2

3

4

5

2004 2005 2006 2007

2003 2004 2005 2006 2007 2003 2004 2005 2006 2007

4

2

0

6

8

10

2003 2004 2005 2006 2007

(^200320042005200620072003200420052006200720032004200520062007)
(a) Consumer price inflation (b) Unemployment rate (c) Wage growth
(d) GDP growth
(g) Money market interest rate (h) Real credit growth (i) Currency depreciation rate
(e) Import price inflation (f) Real exchange rate
Figure 17.11 Dynamic NAM forecasts 2003(1)–2007(3), with end-of-sample for estimation
of parameters in 2002(4)
Actual values are shown by solid lines, and forecasts by dashed lines. The distance between the two dotted
lines represents 70% prediction intervals.
2005(3). Finally, the real exchange rate is systematically overpredicted from 2004(4)
and onwards – the real appreciation of the Norwegian krone is not captured by the
forecasts.
In the following we show how well the NAM forecasts adapt to these location
shifts when we condition on 2004(4) and then 2006(4). In each dynamic forecast
the coefficient estimate is updated. Figure 17.12 shows that the 2005(1)–2007(3)
forecasts for inflation and, in particular, for the interest rate and the real exchange
rate, have improved relative to Figure 17.11. For the other variables there is little
change and, if anything, the forecast failures for the rate of unemployment and for
GDP growth stand out more clearly than before (also showing that there is a knock-
on effect of the high employment forecast on wage inflation). The explanation may
be that NAM is unable to adapt, or that the location shifts of the variables are of
the after-forecast type.