China is in the news; or more accurately, the Chinese currency, is. From Reuters:
A sharp rise in China’s yuan currency might cut the U.S. trade deficit by as much as one third and create enough American jobs to put at least a modest dent in the unemployment rate.
Then again, it may also lead to a destabilizing spike in Chinese unemployment and spark a trade war that drags the global economy back into a deep recession.
I don’t know how the politics surrounding the House bill will unfold. I will also defer on the issue of whether the threat of tariffs would induce Chinese acquiescence or spur retaliation (see Kash Mansouri for discussion of these issues.) I’ll focus on the economic (specifically trade flow) effects of a yuan appreciation. To begin with, it’s useful to think about where we’ve been, and where we’re going.
Figure 1: Log real value of CNY (blue. left axis), 2005=0; up is appreciation of the Chinese currency, and monthly trade balance, in bn USD (red, right axis), and trailing 12 month moving average Chinese trade balance (purple). Note: Real exchange rate uses weighte daverage of swap and official exchange rates for pre-1994. Source: St. Louis Fed FREDII, IMF IFS, ADB ARIC, Fernald, Edison and Loungani (JIMF, 1999), and author’s calculations.
Since over the recent period the real value of the CNY and the trade balance have covaried positively, some might think that real exchange rates don’t matter. However, we know that other variables are relevant, so it is useful to pursue this matter a little more deeply.
It might be true that the a 20% revaluation of the yuan would induce the US-China trade balance effects Bergsten mentions.* In my view, what is important is the impact on the Chinese overall trade balance; in macroeconomics, we typically care about overall trade balances, rather than bilateral (I know politics is different). One way to illustrate this is that given a Chinese revaluation, the US trade deficit might be re-apportioned to other countries. From the article:
“An appreciation of the yuan against the dollar would indeed reduce the U.S. trade deficit with China, but it is unlikely to have a major effect on U.S. job creation,” said Eswar Prasad,** a former International Monetary Fund official who now teaches international trade policy at Cornell University in New York.
The fact that yuan revaluation would not necessarily have a large impact on job creation (for that we need vigorous and immediate expansionary monetary and fiscal policy) does not mean that it wouldn’t be a (very) good thing, however. I think Eswar as well believes that accelerated yuan appreciation would be a good thing, in that it would facilitate global adjustment of current account balances, as well as the transition to a new development model for China. Paul Krugman has forcefully argued that accelerated yuan appreciation would help US employment, based on work by Autor, Dorn and Hanson (2011) (discussed in this post), and I’m willing to be convinced.
Returning to (an utterly conventional) partial equilibrium elasticities approach, what would be the impact on the Chinese trade balance of a yuan appreciation? Many of the studies that have been cited have been hampered by the lack of precise estimates of the relevant trade elasticities. In previous posts   , I have discussed some of these estimates; the import elasticities have been particularly problematic. (On this general topic, see also ) Specifically, in a standard specification, the real exchange rate has the wrong sign.
In a recent paper, Yin-Wong Cheung, Xing Wang (Kevin) Qian and I estimate export and import elasticities using dynamic OLS over a more recent sample spanning the Great Recession. We disaggregate the data, allow for supply effects, account for processing trade, as Yin-Wong Cheung, Eiji Fujii and I did in our 2010 paper. From “Are Chinese Trade Flows Different?”
we have found that for exports, while there is some diversity of responses to income and exchange rate variables, it appears to be the case that Chinese trade flows to accord with priors. Higher rest of world income results in higher Chinese exports, while a stronger yuan results in lower exports. However, the income elasticity is imprecisely estimated, varying widely depending upon the inclusion or exclusion of a linear time trend. In addition, the price elasticity varies widely between goods exported from [state owned enterprises] SOEs, foreign invested firms, and private firms. The latter appear to behave in a more price-sensitive fashion than the other types. As their share of exports continues to rise, one should expect the overall price elasticity to increase, holding all else constant.
We have replicated some of the puzzling results that other researchers have found relating to Chinese import behavior – namely an apparently negative income elasticity. We are tempted to ascribe this result to import substitution taking place as the Chinese economy’s structure alters drastically. However, that remains purely a conjecture.
On the other hand, the fact that disaggregation and the use of proxies for sectoral demands leads to positive coefficients on the activity variables is consistent with the view that rapid structural change has resulted in what appears to be unstable and perverse income and price elasticities at the aggregate level.
Disaggregation, however, only improves the results. We are unable obtain a specification without a large negative price coefficient for manufactured imports, which still constitute about 70% of Chinese imports, except when relative productivity is included. Indeed, inclusion of relative productivity makes the estimates much more plausible, and results in a correctly-signed exchange rate coefficient. To the extent that relative productivity is correlated with supply, this outcome is not entirely surprising.
In the end, Chinese trade flows do seem to respond to economic activity and price variables in the expected manner, at least when the data are sufficiently disaggregated or supply factors incorporated. This means that policymakers in principle had the means by which they could have affected the trade balance in the run-up to the crisis, had they wished.
Our estimated elasticities are obtained using dynamic OLS, which provides long run elasticities. In my next post, I’ll discuss the relationship of these estimates, along with those obtained using an error correction specification, to assessing the exchange rate impact on the Chinese trade balance (as opposed to Chinese exports).
*Disclosure 1: I serve on the academic advisory council of the Peterson Institute for International Economics.
** Disclosure 2: Eswar Prasad is a former coauthor.
This post originally appeared at Econbrowser and is reproduced with permission.