To What Extent Do Exchange Rates and their Volatility Affect Trade?

30
– TO WHAT EXTENT DO EXCHANGE RATES AND THEIR VOLATILITY AFFECT TRADE 
OECD TRADE POLICY WORKING PAPER NO. 119 © OECD 2011 
Annex B. 
 
Econometric Methodology 
Pesaran 
et al.
(2001) suggest an alternative technique
1
, the Auto-Regressive 
Distributed Lag (ARDL) or the bounds test approach to cointegration
2
to investigate the 
relationship between variables. The tests for long run relationship between variables are 
based on standard F-tests. There is no need for pre-unit-root testing. This is one of the 
main advantages of the bounds testing approach which makes it relatively more relevant 
for our topic because the volatility measure could be stationary whereas other variables 
could be non-stationary (Bahmani-Oskooee and Mitra (2008)). Besides, this technique 
generally provides unbiased estimates of the long run model and valid t-statistics even 
when some of the regressors are endogenous (Harris and Sollis (2003)). Inder (1993) and 
Pesaran and Pesaran (1997) have shown that the inclusion of the dynamics may help 
correct the endogeneity bias. Finally, the advantage of this approach is that it allows the 
distinction between short and long run effects. 
The equations of imports and exports of product 
i 
(
i 
stands agriculture or 
manufactured) are modelled as a conditional ARDL-error correction model for each pair 
countries (Euro area with the US and then China; the US with China).
t
t
t
t
country
t
i
k
t
n
k
k
k
t
n
k
k
k
t
country
n
k
k
k
t
i
n
k
k
it
Vol
ER
Y
M
Vol
c
ER
c
Y
c
M
c
c
M




































1
3
1
2
1
,
1
1
,
0
4
0
4
3
0
3
,
2
0
2
,
1
1
1
0
ln
ln
ln
ln
ln
ln
ln
ln
ln
t
t
t
t