Palgrave Handbook of Econometrics: Applied Econometrics

Tommaso Proietti 421

1950 1960 1970 1980 1990 2000

0

1

Monthly GDP and trend

1950 1960 1970 1980 1990 2000

–0.05

0.00

0.05

Monthly output gap, interval estimates

1950 1960 1970 1980 1990 2000

3

4

Industrial production and trend

1950 1960 1970 1980 1990 2000

1.0

1.5

2.0

2.5

Unemployment and NAIRU

1950 1960 1970 1980 1990 2000

–0.25

0.00

0.25

0.50

Cyclical unemployment

1950 1960 1970 1980 1990 2000

0.00

0.01

CPI monthly inflation and trend

Figure 9.9 Monthly multivariate output gap model with temporal aggregation constraints.
Smoothed estimates of monthly GDP, potential output, output gap, NAIRU, cyclical
unemployment and core inflation

significant difference at the end of the sample. Also, enlarging the information set
is beneficial to the reliability of the output gap estimates.
If the model is extended to allow for correlation between the output gap distur-
banceκtand the trend disturbanceηt, as in section 9.2.5, but in a multivariate
set-up, the estimated correlation isˆr=0.10 and does not significantly differ from
zero. In fact, the model with correlated disturbances has a likelihood of 7263.51,
whereas the maximized likelihood of the restricted model (r=0) is 7263.28. Thus,
the LR test ofH 0 :r=0 takes the value 0.459, withp-value 0.50.

9.4 The reliability of the output gap measurement

The reliability of the output gap measurement is the subject of rich debate, and also
has strong implications for optimal monetary policy. Orphanides and van Norden
(2002) and Camba-Méndez and Rodriguez-Palenzuela (2003) discuss the differ-
ent sources of uncertainty and their empirical assessment. The former conclude
that the real-time estimates are unreliable. This conclusion echoes that by Staiger,
Stock and Watson (1997) and Laubach (2001) concerning the NAIRU, obtained
from a variety of methods. Somewhat different conclusions are reached by Planas