Is there a J-curve for Azerbaijan? New evidence from industry-level analysis

Abstract

This paper estimates the J-curve for Azerbaijan using quarterly industry-level data over the 2000–2009 period. Empirical results show that in 3 of the 10 strategic industries there is strong evidence for the fulfilment of the Marshall–Lerner condition, as the trade balance improves in the long run in reaction to a currency devaluation. In most industries the J-curve pattern is observed in the short run. All 10 cases exhibit long-run cointegration and are stable according to the CUSUM and CUSUMSQ stability tests. These findings are largely consistent with the existing literature on the Azerbaijani J-curve and carry important policy implications.

Keywords:

• J-curve
• ARDL regression
• Azerbaijan

JEL Classification:

• C22
• F14
• F3

Notes

^1. This is also the reason why we cannot use bilateral exchange rates vis-à-vis specific trade partners, because partner-specific non-oil trade information is not available.

^2. We thank the anonymous referee for pointing this out.

^3. This choice was restricted mainly by our small sample size.