The Effect of Exchange Rates on Chinese Trade: A Dual Margin Approach

ABSTRACT

Previous studies investigating the effect of exchange rate changes on a country’s exports have found little evidence that exchange rates matter. This “Exchange Rate Disconnect Puzzle” may stem from the fact that studies have mostly focused on aggregate data. Using HS-6 digit product-level data for Chinese exports, we analyze the effect of real exchange rate (RER) as well as the volatility of RER of the Chinese RMB. By decomposing China’s exports into its “extensive” and “intensive margins,” we find that RER volatility significantly impacts Chinese exports via both the margins. RER volatility increases the uncertainty and deters new firms from entering the market. As less firms operate, the export share of the existing firms increase. The overall effect of this volatility is slightly positive. We find that these effects are dominant for the minor trading partners of China compared to its major trading partners. We find weak evidence that RER depreciation affects China’s exports via the extensive margin.

KEY WORDS:

• Chinese trade
• extensive margin and intensive margin
• real exchange rate
• volatility

JEL:

• F14
• F31

Acknowledgment

We thank the seminar participants at the WEAI 2015 conference. All remaining errors are ours.

Supplementary material

Supplementary data can be accessed here.

Notes

1. China’s current account surplus reached more than 10 per cent of GDP in 2007, prior to the global financial crisis.

2. The depreciation of the Chinese currency and its effect on Chinese exports has been in recent news with the escalating trade war between United States and China. Many believe that China has its currency to use as a weapon in the trade war with United States. Since the Trump administration’s imposition of tariffs on Chinese goods, the RMB has lost about 4%. This weaker currency has been a sore point for a number of U.S. administrations, which have blasted China for allowing the RMB to weaken to help exports.

3. Typically, the effects of currency depreciation on a country’s trade balance follows a “J-curve” effect. This suggests that trade balance is negative in the short-run following a currency depreciation and only becomes positive in the long-run when the Marshall–Lerner Condition is satisfied.

4. For other work in this area, see the survey by McKenzie (1999) who investigates the link between exchange rate uncertainty and trade flows and concludes that exchange rate volatility may impact differently on different markets and proposes that further tests using export market specific data be conducted.

5. Some recent studies cast doubt over the conflicting evidence. See Dekle and Ryoo (2007) and Leigh et al. (2015) for further details.

6. For recent theoretical literature on trade margins, see Hummels and Klenow (2005), Chaney (2008) and Helpman, Melitz, and Rubinstein (2008) among others.

7. Their finding of new firms exporting small amounts and facing high export failure rates is also similar to the findings of Besedes and Prusa (2011) who found that most export relations are very short lived.

8. Kang and Lynn (2009) stress on the importance of extensive margin of international trade and believes that the export price index will be underestimated without taking extensive margin into account.

9. These measures of extensive and intensive margins have been used in the literature by Bergin and Lin (2009), Lin (2012) and Dutt, Mihov, and Zandt (2013) among others.

10. The extensive margin of exports to any country should lie between 0 and 1, both inclusive. The lower bound of the intensive margin is zero, but it can be greater than 1.

11. Adding all other explanatory variables, the sample reduces to 165 countries.

12. Some researchers have used variants of ARCH models as a proxy for exchange rate variability, while very few researchers have used daily spot exchange rates to compute 1-month ahead exchange rate volatility via a method based on Merton (1980).

13. These figures are based on data from World Bank’s World Development Indicators.

14. For details, see the derivation of the gravity equation in the theoretical model presented in the Appendix.

15. This result is in line with a similar finding by Lin (2012).

16. We thank one of the referees for suggesting this explanation.

17. Melitz (2003), Chaney (2008), Redding (2011), and Arkolakis, Costinot, and Rodriguez-Clare (2012) suggest that the intensive margin elasticity could be higher than the extensive margin elasticity, while Foster, Poeschl, and Stehrer (2011) find conflicting empirical result for this effect.

18. ${\tau }_{ij}>1$ for any $i\ne j$ and ${\tau }_{ij}=1$ for $i=j$.

19. Here, ${\theta }_j^h$ is similar to the multilateral resistance variable of Anderson and van Wincoop (2003).