correction specification, equation (5), includes F and, in addition, other exogenous variables, X,
which includes the nominal effective real exchange rate, NEER, and the ratio of domestic credit
to GDP, DCRE, to capture the effects of nominal devaluation and expansionary macroeconomic
policy, respectively, on the short run movements of real exchange rate. A time trend, TREND, is
also included. As Dufrenot and Egert (2005) and Taylor and Sarno (2001, p.157) note, the time
trend is intended to capture several factors that may cause an appreciation of the real exchange
rate in addition to productivity changes, including increased demand for tradable goods relative
to non-tradables as income increases during transition, changes in the composition of the CPI
which may cause the CPI to increase and the real exchange rate to appreciate, and productivity
changes not captured elsewhere. 19
The sign beneath each variable in (6) indicates the expected sign of the partial derivative in
both equations (4), the steady state relationship, and (5), the short run error correction
specification. The theoretical model indicate that an increase in the terms of trade will raise
purchasing power and so, the domestic demand for all goods increases. Under the small country
assumption, the price of the traded goods remains constant, but the price of non-traded goods
rises. Thus, the equilibrium real exchange rate increases. This is the income effect. However, an
increase in the terms of trade will also have substitution effects on both demand and supply
sides. Consumers will shift from the consumption of exportables and non-traded goods to the
consumption of importables. This increases imports, and lowers the price of non-traded goods,
causing a fall in the real exchange rate. On the other hand, producers will increase production of
exportables and decrease production of non-traded goods. This will raise the real exchange rate.
19 Dufrenot and Egert (2005) estimate a slightly different specification for the real exchange rate in the Czech
Republic, Hungary, Poland, Slovakia and Slovenia in a two equation structural vector autoregressive model in order
to determine whether the trend appreciations in the real exchange rate exhibited in transition economies were
movements toward the equilibrium rate or there was still potential disequilibrium.
14