Manufacturing export prices for the G7.
Barrell, R.J.
MANUFACTURING EXPORT PRICES FOR THE G7
The Institute World Model, GEM, has equations for manufacturing
export prices for each of the G7 countries. The long-run solutions to
the equations are all of the simple form:-- PXG =
[gamma].sub.0.WPXG[gmma].sup.1.(PD/RX)[gmma].sup.2.sup.e..sup.W where
PXG: manufacturing export prices in dollars WPXG: competitors export
prices in dollars PD: domestic wholesale prices in domestic currency RX:
domestic currency units per dollar
These equations have recently been re-estimated over the period
1965Q1 to 1987Q3. In that estimation we imposed the condition that the
equation is homogenous of degree one in its arguments, WPXG and PD/RX,
by assuming that [gmma].sub.1 + [gmma].sub.2 = 1, and we included
dynamic elements in all the variables. The general form we estimated
was (lower case indicates logs). and we produced a parsimonious form by
sequentially eliminating the dynamic elements.
The long-run characterics of the equation are given in table 1. The
higher the weight on domestic prices the less influence there is on
export prices from overseas prices. The US equation is very simple, in
that only domestic prices affect export prices, reflecting both the size
of US industrial production and the small proportion of the production
which is exported. Japan and Germany also displayed a rather small role
for world prices in the long-run solution, which is not surprising given
the size of their industrial sectors. France appears to be the most
open of these seven economies. The high weight on Canadian domestic
prices may reflect the direct influence of US wholesale prices on
Canadian prices, rather than indicate a low degree of openness. The law
of one price would suggest that, in the long run, only world prices
would determine export prices. None of the countries studied displayed
this property, reflecting the heterogenous nature of the goods traded
and differing degrees of market power. In no case was the law of one
price form of the equation statistically better than the form chosen.
The dynamic properties of the equations can be summarised by the
mean lags reported in table 2. We did not choose to impose the same lag
structure on domestic prices as on exchange rates. Exchange rate
changes may initially be seen as less permanent than changes in domestic
prices, and hence a permanent change in exchange rates may take longer
to be embedded in prices than does a change in domestic prices. This
does indeed appear to be the case for Germany, France, italy and the UK.
Response times to changes in world prices appear to be lowest in the
open economies of Italy and the UK, and longest in the relatively closed
German economy. We have also not chosen to impose homogeneity of degree
one on the coefficients of the dynamic part of the equations, although
it would appear to be a valid restriction for all economies except
Japan, where a change in world inflation rates does seem to have a
long-run effect on Japanese export prices.
