Yen) in their export demand equation.

To measure volatility they used the standard

deviation of the growth rates of real exchange rates (12-months moving-average) and

the coefficient of variation of the real exchange rate. The authors found evidence of

both positive and negative effects for the 1975-1998 period. However, they did not

report any formal stationarity inspection through unit root tests prior OLS estimation.

Because it is not known whether the variables are stationary (equal variance) or non-

stationary, and if the latter is the case, then estimation results are spurious (Granger

and Newbold (1974)). Homoscedastic variance is a necessary assumption for OLS to

yield reliable results.

With the development of new theories and methodologies in time-series

econometrics such as co-integration and error correction models (Engle and Granger,

1987; and Johansen, 1988 and 1991), economists started to look at the long-run

relationships between exchange rate volatility and export flows (e.g., Erdal et al.,

(2012); and Arize et al., (2000)). Arize et al. (2000) found negative and significant

relationship in both short and long-run effects of real effective exchange rate

volatility on export flows from 13 developing countries (Ecuador, Mexico, Mauritius,

Morocco, Malaysia, Malawi, Indonesia, Korea, Philippines, Sri Lanka, Taiwan,

Thailand, and Tunisia) for the 1973-1996 period. Using the same approach Arize et

al. (2008) also found negative effects for Latin American countries (Bolivia,

Colombia, Costa Rica, The Dominican Republic, Ecuador, Honduras, Peru and

Venezuela). Vergil (2002) examined the effect of exchange rate volatility on Turkish

bilateral trade flows to the United States, Germany, France, and Italy. The authors

used moving standard deviation of rates of change of exchange rates as measure of

21