Exchange rate volatility and Turkish industry-level export: Panel cointegration analysis

Abstract

The purpose of this article is to investigate the impact of the exchange rate volatility on Turkey's export. To this end, the panel cointegration analysis is applied to the data from Turkey's top 20 export industries to major 20 trading partners for the period 1980–2009. Special attention is paid to test for whether employment of country-level trade data instead of industry-level data is subject to the aggregation bias problem in the estimation of long-run cointegration parameters. The results indicate that employing country-level trade data suffers from the aggregation bias in estimating the cointegration parameters for the level of exchange rate and for the exchange rate volatility. The findings imply that (i) the impact of the exchange rate volatility on Turkish exports differs across industries, (ii) Turkey benefits from the depreciation of Turkish lira, and(iii) the foreign income plays a key role in determining the Turkish industry-level exports. The findings increase our insights to explain therecent dynamics of Turkish exports and provide some policy implications.

Keywords:

• exchange rate volatility
• industry-level export
• Turkey
• panel cointegration

JEL Classifications:

• F14
• F31
• C33

Acknowledgements

I would like to thank anonymous reviewers and the editor Professor David Giles for their constructive and valuable comments that help me to improve the paper. Any remaining errors are the author own responsibility.

Notes

1. An interested reader is referred to McKenzie (1999), Ozturk (2006), Bahmani-Oskooee and Hegerty (2007), and Hall et al. (2010) for comprehensive reviews of the literature on the exchange rate volatility and trade linkage.

2. See first and third columns of Table 3 for Turkey's top 20 export industries and major trading partners, respectively.

3. I refer an interested reader to Asikoglu and Uctum (1992) for a broad overview of Turkish exchange rate policies during the 1980–1990 and CBRT (2002) for an overview of the liberalization process.

4. The conversion rates: 13.7603 for Austria, 40.3399 for Belgium, 5.94573 for Finland, 6.55957 for France, 1.95583 for Germany, 340.75 for Greece, 0.787564 for Ireland, 1936.27 for Italy, 2.20371 for Netherlands, 200.482 for Portugal, and 166.386 for Spain.

5. See Bahmani-Oskooee and Hegerty (2007) for a wide range of the proxies in measuring exchange rate volatility.

6. In order to save space, results for GARCH(1,1) estimations are not reported here but available upon request.

7. It is important to note here that the first generation non-stationary panel methods assume independency among cross-sectional units. The second generation panel unit root and cointegration tests take into account cross-sectional dependency; however, they require balanced panel data sets (Demetriades and James 2011). Since this study employs unbalanced panel data set, the first generation panel methods are more suitable than the second generation panel tests.

8. In order to save space, the results from the unit root tests of Hadri (2000), Levin, Lin, and Chu (2002), and Im, Pesaran, and Shin (2003) are not reported here, but are available upon request.