TO WHAT EXTENT DO EXCHANGE RATES AND THEIR VOLATILITY AFFECT TRADE –
43
OECD TRADE POLICY WORKING PAPER NO. 119 © OECD 2011
Annex G.
Diagnostic Tests
Some of the diagnostic tests undertaken are reported below. According to
the adjusted
R
2
, the explanatory power of our estimated models is satisfactory. Additionally, CUSUM
and CUSUMSQ tests mostly support stability of the short-run and the long-run
coefficient estimates. (Graphical presentations of these test results have been not reported
for reasons of space but they are available upon request.) Third,
the LM test for serial
correlation is significant only in four cases out of the 24 models. Finally, Ramsey’s Reset
test clearly indicates that all 24 models are well specified.
Table G.1. Diagnostic tests with GARCH volatility measure
Import-value model
2
R
CUSUM
CUSUMQ
LM
a
RESET
b
Euro Area / United States
Agriculture sector
Non-agriculture sector
0.77
0.65
Stable
Stable
Unstable
Stable
30.80
28.96
0.22
0.00
Euro Area /China
Agriculture sector
Non-agriculture sector
0.44
0.53
Stable
Stable
Stable
Stable
19.07
21.58
0.14
0.47
United States /China
Agriculture sector
Non-agriculture sector
0.42
0.64
Stable
Stable
Stable
Stable
19.30
20.73
0.09
0.73
Export-value model
2
R
CUSUM
CUSUMQ
LM
RESET
Euro Area / United States
Agriculture sector
Non-agriculture sector
0.76
0.75
Stable
Stable
Stable
Stable
21.76
20.86
0.62
0.76
Euro Area /China
Agriculture sector
Non-agriculture sector
0.47
0.69
Stable
Stable
Unstable
Stable
20.68
34.02
0.68
2.37
United States /China
Agriculture sector
Non-agriculture sector
0.48
0.36
Stable
Stable
Stable
Stable
15.13
21.67
1.53
1.32
44
– COMPARATIVE ADVANTAGE AND TRADE PERFORMANCE: POLICY IMPLICATIONS
OECD TRADE POLICY WORKING PAPER NO. 119 © OECD 2011
Table G.2. Diagnostic tests with five-year moving standard deviation volatility measure
Import-value model
2
R
CUSUM
CUSUMQ
LM
a
RESET
b
Euro Area / United States
Agriculture sector
Non-agriculture sector
0.76
0.69
Stable
Stable
Stable
Unstable
22.88
25.57
0.00
0.81
Euro Area / China
Agriculture sector
Non-agriculture
sector
0.46
0.61
Stable
Stable
Stable
Stable
16.43
19.05
0.17
1.74
United States / China
Agriculture sector
Non- agriculture sector
0.53
0.65
Stable
Stable
Stable
Stable
8.54
21.63
0.52
1.15
Export-value model
2
R
CUSUM
CUSUMQ
LM
RESET
Euro Area / United States
Agriculture sector
Non- agriculture sector
0.71
0.75
Stable
Stable
Stable
Stable
27.54
18.61
0.02
0.04
Euro Area / China
Agriculture sector
Non-agriculture sector
0.47
0.72
Stable
Stable
Unstable
Unstable
17.91
13.07
0.64
0.00
United States / China
Agriculture sector
Non- agriculture sector
0.51
0.41
Stable
Stable
Unstable
Stable
20.16
20.45
0.95
1.15
a. The Lagrange Multiplier Test (LM) of residual correlation is distributed as
2
with 12 degrees of freedom. At the
5% (1%) level of significance, its critical value is 21.03 (26.22)
b.
Ramsey’s Reset test for functional misspecification is distributed as
2
with one degree of freedom. Its critical
at 5% significance level is 3.84.