3. Estimation results
The equations are estimated with fixed effects.7 As all equations include lagged dependent
variables, the estimators could be biased as the lagged dependent variable is correlated with
the fixed effects. It has been proposed in the literature to use the Arellano Bond estimation
technique which is based on a GMM estimation of the differenced equation. However, Judson
and Owen (1999) have shown that the bias is small when the time dimension is large relative
to the cross-sectional dimension. Indeed, they find a negligible bias for a time dimension of
30 or larger. Given that our time dimension is mostly around 50 quarters or more, with some
exceptions in Portugal, the Netherlands and Luxemburg (only for energy and intermediate
goods producer prices) we do not use the Arellano-Bond estimator.
Despite a significant number of exogenous variables, the estimation of the pass-through
should not be affected by strong co-movements between the exogenous variables as there is
relatively little contemporaneous correlation among them (see Table 2). No correlation
coefficient is above 0.5. The highest correlation exists between energy and food commodity
prices (0.41), followed by the correlation between food and industrial raw material
commodity prices (0.25). Both correlations could reflect a third driving factor such as global
demand and/or the high energy content for food and industrial raw material production.
Table 2 Contemporaneous correlation across exogenous variables
|
EXTRAj | |||||||||
|
NEER |
COMENE COMFD |
COMIRM |
VAT |
ULC |
YGAP |
OPEN |
ENETAX | ||
|
NEER |
1.00 | ||||||||
|
COMENE |
-0.05 |
1.00 | |||||||
|
COMFD |
0.13 |
0.41 |
1.00 | ||||||
|
COMIRM |
0.03 |
-0.03 |
0.25 |
1.00 | |||||
|
VAT |
0.01 |
-0.08 |
-0.03 |
-0.04 |
1.00 | ||||
|
ULC |
0.10 |
-0.07 |
-0.14 |
-0.05 |
0.03 |
1.00 | |||
|
YGAP |
-0.09 |
0.17 |
0.24 |
0.03 |
-0.05 |
-0.36 |
1.00 | ||
|
EXTRAJOPEN |
0.00 |
0.02 |
0.08 |
0.04 |
-0.15 |
-0.05 |
-0.05 |
1.00 | |
|
ENETAX |
-0.05 |
0.01 |
0.09 |
0.10 |
0.06 |
-0.03 |
-0.01 |
-0.06 |
1.00 |
NEER: nominal effective exchange rate of the euro; COMENE: energy commodity prices in USD; COMFD: food
commodity prices in USD; COMIRM: industrial raw material prices in USD; VAT: value added tax; ULC: unit
labour costs; YGAP: output gap; EXTRA_OPEN: extra-euro area trade openness; ENETAX: energy taxes.
Table 3 shows which of the theoretically possible causal relationships (shaded area) in the
estimated pricing chain have been found to be significant (the regression results can be found
7 Random effects estimations yielded similar results.
И ECB
Working Paper Series No 1104
November 2009