Real exchange rate determinants in transition economies
MetadataShow full item record
The real exchange rate being the most important relative price in international economics plays a critical role in determining the competitiveness, resource allocation, the direction of international trade, and the economic growth and development in a particular country. Political factors are often cited in the theoretical literature as potential determinants of foreign exchange volatility. However, these factors are seldom captured in empirical exchange rate models, especially in transition and developing countries. Armenia witnessed an impressive economic growth during its transition from the Soviet centrally planned to the market economy. By mid-2000s, Armenia was able to achieve a sustained economic growth, which was accompanied by drastic appreciation of Armenia’s national currency by 46% between 2003 and 2007. Rapid appreciation triggered alarms and gave rise to speculative theories of government manipulation to pocket hard currency and benefit government-connected importers. Central Bank denied all wrongdoing and insisted that drastic increases in dollar remittances and economic growth were major cause. Thus, scientifically robust explanations are timely and crucial to avoid further speculation and manipulations for political gains. This study utilizes multiple econometric estimation approaches (VAR, VECM and ARDL) to analyze real exchange rate dynamics in relation to economic fundamentals. We then incorporate several political risk indicators in the analysis to see if they improve the overall performance of the real exchange rate models and inquire if changes in the political climate have affected the real exchange rate in Armenia. Finally, we evaluate the out-of-sample forecasting power of these models to (1) make an inference on the overall effectiveness of these models to forecast real exchange rate in a transition country; (2) examine if accounting for the political climate and investment risk helps to improve the forecasting power; and (3) see if any of these models perform better than the simple random walk as argued in Meese and Rogoff (1983a). Results provide strong indications that the real exchange rate dynamics over the study period were driven by economic developments and weigh against the claim that the government and the Central Bank directly manipulated the exchange rate. Evaluation of alternative estimation approaches based on their out-of-sample forecasting performance provide a strong support for a more recent bounds testing and ARDL estimation approaches over the traditional VAR and VECM in out-of-sample forecasting performance. Furthermore, results indicate that the ARDL models (both, with and without political risk) perform slightly better in out-of-sample forecasting as compared to the random walk. Our findings empirically confirm the theoretical findings of Pesaran and Shin (1998) and Pesaran et al. (2001) for this analysis. Even though the initial analysis showed that the political risk indicators are often correlated with the main political and economic events in Armenia, our results suggest very weak or no effect of the political risk factors on Armenia’s real exchange rate. This suggests a minor role for political risk in the decisions of major foreign investors in Armenia, which may in part be explained by a different investment decision-making rationale for Diaspora represented investors.