Abstract
This article investigates how the response to devaluation of trade balance is affected, compared to J-curve hypothesis, by the presence of imported inputs in the production of exports. Using first the Almon lag technique and then the cointegration and the generalized impulse response function analysis, the J-curve effect is examined in two sectors of Turkish economy (manufacturing and mining), which use imported inputs at different rates. Based on the data covering the period from the first quarter of 1986 to the third quarter of 1998, our results indicate that in neither sector J-curve exists and that the violation of the J-curve effect is more severe in the sector with higher import content
Acknowlegements
We are indebted to one anonymous referee of this journal for his useful comments. The usual disclaimer applies.
Notes
1For a more detailed review of the relevant studies, see Bahmani-Oskooee and Ratha (Citation2004a).
2These ratios are calculated from Table A: Aggregated Input-Output Table of Turkey and Table 9: Input-Output Table for Imports in Input-Output Structure of Turkish Economy 1998.
3Data was available only at yearly basis for 1986. This yearly value has been converted into quarterly values, using quarterly industrial production index as weights. The source for this index is also Turkish Central Bank (www.tcmb.gov.tr).
4The share of each country in Turkey's total trade out of these 14 countries in order of importance is Germany: 0.238, Italy: 0.128, UK: 0.118, US: 0.111, France: 0.109, Spain: 0.067, Netherlands: 0.048, Switzerland: 0.037, Belgium: 0.035, Israel: 0.025, Japan: 0.025, South Korea: 0.023, Sweden: 0.018, Austria: 0.018.
5To be consistent with the literature that uses vector error correction modelling, money variables included in the Almon lag technique are not included in this part.
6Expected signs of the coefficients of the model variables are explained in the previous method.
7VECM estimation results are not provided here but available upon request.