March 28, 2006
What would be the important effects?
With the impending decision by the Treasury regarding whether the Rmb has been manipulated, and the surging protectionist pressures in the Congress (see here), it might be useful to think about the important effects of a Chinese revaluation, were it to occur.
Contrary to popular -- although not unanimous -- opinion, I do not believe the biggest effect of the revaluation will be on the amount of imports into the U.S., or even the amount of Chinese imports. This view was echoed in a recent FT article.
There are two parts to my reasoning. First, the relevant price elasticities are quite small. In fact, in one study, the relevant estimated elasticity for the 1994-2001 period is not statistically significant (see also Mann and Plueck's estimates in Table 6). One set of elasticities used by Goldman Sachs indicates a Chinese import price elasticity of 0.2 and an export price elasticity of 0.5 (see this CRS report). The sum of these elasticities is less than one, indicating that the Marshall-Lerner conditions do not hold. Now, it is unclear (to me) how these estimates were obtained, and what the extent of uncertainty surrounding these estimates are. But I suspect the presumption that the elasticities are low make sense, given that there is a large imported component in Chinese exports (see some slightly older estimates by K.C. Fung).
Second, to the extent that the U.S. trade imbalance is driven by U.S. savings and investment patterns, a Chinese revaluation (holding all else constant) would merely re-allocate the U.S. trade deficit to India, Thailand, Malaysia, Korea,... (The graph below is helpful in that regard, even if it is slightly out of date -- it shows the fact that China has run a trade deficit with the rest-of-the-world ex. US).
Source: MarketThoughts.com, "China - Basic Background and Current Issues, Part III."
Addition on 3/29/06: Brad Setser is right -- here is an updated CA picture for China, based on IMF WEO data.
Source: IMF, WEO, Sep. 2005.
This is not to say that accelerated Chinese appreciation would have zero impact. In fact, if other East Asian currencies appreciate in line with the Rmb (i.e., if China is a first-mover), then this will mean that acquistion of U.S. Treasury securities will be moderated (and as we all know from Brad Setser's work, Chinese acquisitions alone are substantial), U.S. interest rates rise, further depressing consumption and asset price behavior in the U.S. This is where China's exchange rate policy is important -- not in changing the relative price of Chinese goods, but rather in altering asset prices and hence capital flows.
Final point: As I have stressed elsewhere, Chinese revaluation in this context is an important adjunct to global rebalancing; but it is not a substitute for fiscal consolidation in the U.S. Something to remember, as the public debate centers on China, and ignores the impact of the burgeoning U.S. budget deficit and the rising share of the trade deficit arising from high enery prices and energy import dependence.
Posted by Menzie Chinn at March 28, 2006 07:03 PMdigg this | reddit
Listed below are links to weblogs that reference Chinese revaluation:
» US trade deficit dilemma from Financist
If Federal government doesn't change the reckless fiscal policy, there won't been such a huge deficit to be funded by foreign capital. Say US don't want to tango with China now. It doesn't mean that US will stop tango. Who's gonna finance the overspe... [Read More]
Tracked on April 9, 2006 05:16 AM
» Blogosfera from De Gustibus Non Est Disputandum
Rápidas * Outra boa pergunta. * Starbucks (infelizmente, não a da nova Gallactica). * Oriente Médio liberal? * Portugueses com Escolha Pública e contra Austríacos? Sim. * O que os outros cientistas sociais acham disso, eu não sei. Mas economistas... [Read More]
Tracked on April 9, 2006 02:59 PM
Thanks for the post. I think you presented well the degree of uncertainty ahead for the necessary task of global rebalancing. In my opinion, the process will involve pain but ignoring the problem will only lead to the liklihood of a financial meltdown and a harder landing.
I have not seen any data on the elasticity of external demand for treasury securities as a function of the interest rate or the relative value of the currencies. The Chinese and other export oriented nations have been buying this debt for the past few years without regard to potential capital losses. Although rates of the longer term securities has not changed significantly over the course of the external purchases, the potential for huge capital losses from increasing rates has always been a distinct possibility. The current holders of this debt have an incentive to continue lending at lower rates because of the capital loss they will experience on the debt they already hold. In addition, they have an enormous incentive to keep the American consumer spending. I suspect that the external lending market is also inelastic and longer term interest rates may continue to be "sticky" .
Posted by: anon at March 28, 2006 08:34 PM
The reduction in profitability of the already unprofitable Chinese manufacturing sector isn't an insignificant effect. Yuan revaluation could trigger the manifestation of massive Chinese misallocation of resources.
Posted by: algernon at March 28, 2006 08:37 PM
Menzie -- love your post, but not your graph. It doesn't include 2005, and, as we all know, China's trade surplus (global) balooned to $100b (literally, almost off the chart, on its current scale) in 05. I think that changes the trend line significantly as well. In my view (probably a bit different than yours), the devaluation v. the Euro from 02-04 (And failure to appreciate v. the $ as productivity rose) had a significant impact on the underlying trade balance, but that effect was masked by imports associated with the investment boom/ negative terms of trade shocks in 03/04. in 05, the import growth briefly slowed, and boom ... $100b surpluses. Zhou (China's CB governor) now says getting back to your long-term trend line will take a couple of years in the best of circumstances.
Posted by: brad setser at March 28, 2006 08:58 PM
Menzie -- should have added that I completely agree with your assessment of how revaluation influences the US current account balance, and apologies for my somewhat snippy tone (particularly in light of the qualification attached to your graph). the notion that China doesn't have a global surplus b/c it its bilateral surplus with the US (And now europe) is offset elsewhere is one of my pet peeves. Sort of true til recently, but no longer true.
Posted by: brads at March 28, 2006 09:06 PM
A significant reval would definitely effect overall US CA balance... eventually.
(1) There are a lot of products coming in from China that are marginally priced now... I know Chinese exporters who constantly complain about how tight pricing is in America. If the RMB strengthens these programs would go somewhere else.
(2) Many of the 'somewhere else' options are keeping their currency low too (Setser's talked about this a ton)... It would allow some of the other CBs a little room to ease up. As with China there are plenty of marginally priced products coming in from this part of the world too that now would either be made domestically (or at least in NAFTA Zone)... or kiboshed and not compete against existing domestic programs.
Combine effects from (1) & (2) and you probably will have some measurable 're-balancing'.
One caveat - it will take YEARS even after the reval takes effect. Most of these imports (if I understand things correctly) are coming from western MNCs either making the stuff directly in Asia OR contracting with Asian third parties to make stuff for them. I work with these kinds of programs (though with domestic US firms)... these are not 'spot price' contracts. Think MANY years & appropriately hedged long term agreements.
If I was going to import an engine from say Shanghai Auto for my platform... it wouldn't be for just 'one year'... I'd have started design years ago... planned the facility & anticipated multiple scenarios including currency... hedged... built the plant & then expected many years of shipments within a tight cost range. So even if the currency revaled I wouldn't see the pressures until the contracts run out.
Walmart, Target, US OEMs buying components or even finished product... do all of these things.
Eventually however those contracts run out... Plus new programs come up for initial evaluation evereyday... Both of these situations would have to be re-evaluated in light of the new currency situations... so there would be an eventual effect. It just might take a WHOLE LOT longer than anyone realizes.
Plus these increased 'costs' would show up as 'inflation' - think of it as built up or latent inflation - and would probably be significant unless the dollar strengthened again vis-a-vis the RMB - ie revaled back. I don't think the FED could do a lot about this inflation except strengthen the dollar via more rate increases (keep the flow of funds coming in to support the dollar drain). Maybe that is what is going on now & we'd have to get inside their heads to know.
That's my 'not-very-much-of-an-economist-but-working-in-the-salt-mine' take on the thing. Love to hear your critique.
Posted by: dryfly at March 29, 2006 06:14 AM
It seems to me that U.S. demands for a yuan revaluation amount to a partial renegging on our debts. From the Chinese point of view it has to look like a devaluation of the dollar. As the main world reserve currency we have had the luxury of denominating our debt in our own currency. However, think if back in the 80's or 90's Brazil's debt were denominated in their currency and they decided to devalue by 30%, how happy would Citigroup be? We are demanding that China take a capital loss in the tens of billions of $.
Posted by: Dirk van Dijk at March 29, 2006 10:38 AM
Brad Setser: Point well taken; I was lazy and did not retrieve the best graph. I have now added another graph to the post plotting the 2005 estimates and 2006 projections from the September IMF World Economic Outlook. The present Chinese CA surplus is now large, expressed either in USD terms or percentage of Chinese GDP.
Anon: An elasticity of Treasury security purchases with respect to interest differentials would be even less "structural" than the usual elasticity, to the extent that it includes central bank actions.
dryfly: I agree that if China moves, reserve accumulation will likely decline. That indeed is the key reason I think Chinese Rmb flexibility is crucial to accelerating global rebalancing (as I argue in the Council on Foreign Relations report and in the post).
I also concur that the adjustment from the Chinese side (in terms of factories, etc.) could take a long time. But for me, the important adjustments are: decline in purchases of Treasury securities, rise in U.S. long term rates, stabilization in housing market prices and decline in consumption growth, leading finally to decline in U.S. imports. This process takes somewhat less time.
algernon: You may be right that even a modest Rmb appreciation will induce massive reallocations in China. But most of the evidence I've heard is anecdotal, and not supported by the econometrics (admittedly applied to imperfect data).
Posted by: menzie chinn at March 29, 2006 10:42 AM
How is your argument consistent with the balance of payments identity in the short run? Today, Chinese exporters earn dollars, and they can convert the dollars into RMB at the PBoC (indirectly). If they continue to earn the same number of dollars (because of very low elasticities), then they will continue needing to convert them, and if the PBoC doesn't do the conversion, the private sector will have to pick up the slack. The effect is a weaker dollar but identical demand for dollar-denominated assets, since the private sector agents who buy the dollars will have to invest those dollars in dollar-denominated assets. How can there be any effect on net (private+public) asset flows until there is an effect on trade flows? You seem to be arguing that the revaluation eventually affects trade flows by weakening the US economy, but the mechanism for that weakening is a short-run impact on asset flows, which, as far as I can tell, is inconsistent with the absence of a short-run effect on trade flows.
Posted by: knzn at March 29, 2006 03:30 PM
I think the statement
it shows the fact that China has run a trade deficit with the rest-of-the-world ex. US).
is a little misleading. According to this CRS report China runs a trade deficit with a very small number of countries: Germany (heavy equipment), Taiwan and South Korea (mostly electrical equipment), petroleum producing countries. It emphatically does not run a trade deficit with the rest of the world except the U. S.. It runs a trade deficit with the entire world except for a few favored trading partners. I don't know what the precise implications of that might be but it seems to me that considering the results excluding the few rather than in aggregate would be an interesting exercise.
Posted by: Dave Schuler at March 30, 2006 06:44 AM
Dave Schuler: Interesting observation, although I don't know if it changes my view of the situation. I'm assuming you're using the left side (partner country data) of Table A45 in the CRS report. If you look at the right side of Table A45, you'll see more countries that China is running deficits with. Which one is more reliable. As this Fed working paper (Schindler and Beckett, IFDP831, April 2005) notes, the true trade flows are probably bracketed by the Chinese and partner country estimates, at least for the US. That is because of the effect of re-exports through Hong Kong, SAR. Hence, in this regard, I agree with the assessment of the 2004 Economic Report of the President in attributing a large share of the increase in U.S. imports from China to a re-allocation of imports from other East Asian countries.
Posted by: menzie chinn at March 30, 2006 08:37 AM
knzn: Good point. Here's my take. Let the balance of payments be:
CA + KF + ORT == 0
where CA is current account, KF the private financial account (formerly known as the private "capital account"), and ORT as official reserves transactions. The "==" indicates "identically equal to" or true by identity.
Currently CA > 0, KF + ORT is less than 0 primarily because of ORT being less than 0 (i.e., reserve accumulation due to forex intervention). Assume ORT goes to 0 as an extreme example. KF might pick up the slack so CA stays unchanged, consistent with the view that price elasticities are low and all else held constant. But as ORT changes, interest rates and other asset prices also change, so that the ceteris paribus condition might not be appropriate. In particular, as Chinese ORT goes to zero, US interest rates rise (and with it world interest rates). You are right that in some models, if the Chinese private sector picks up the slack completely in purchasing US Treasuries, there will be no change. But there is no guarantee that the Chinese private sector will replicate what the PBoC has been doing (especially since there are extant capital controls). If the composition of purchases are different between private and official sector, then that also induces a differential effect (assuming that different assets are imperfectly substitable).
Posted by: menzie chinn at March 30, 2006 08:39 AM
Let's separate the issues of the general level of demand for US assets and the composition of that demand. Regarding the composition issue, I can see that, since private investors may have different preferences than central banks, a revaluation might result in a shift. For example, there might be a shift in demand from bonds to stocks, causing both interest rates and stock prices to rise in the US. In this case, the effect on consumption would be ambiguous. It isn't obvious to me in general that the likely shifts would tend to favor lower consumption.
Regarding the level of demand issue, I think it is unambiguous in the short run, especially if we assume a J-curve effect. There will initially be just as many (in fact, more) dollars going abroad, and those dollars will have to find a home. The private sector will have to pick up the slack, whether or not it initially wants to do so. The problem will not be an excess supply of US assets but an excess supply of dollars, and the value of the dollar will adjust to remove that excess supply. Once the exchange rate adjusts, we have the same number of dollars (actually more dollars) chasing the same set of assets, and it's hard for me to see how the new equilibrium could have a lower price for the abstract US asset.
After the Plaza Accord, there was a dramatic increase in the prices of US assets. I realize it's not quite the same thing. In that case the central banks weren't ending a policy of buying dollars directly; rather, they were ending a policy of maintaining interest rate differentials that had induced the private sector to buy dollars. I believe the same logic should apply, however. After Plaza, investors should have found US assets less attractive, but the prices of those assets went up. It seems to me that, given the J-curve, foreign investors were essentially a captive audience. They could reduce their bid for dollars, but eventually, at some price, someone had to accept those dollars and invest them in the US, producing a continued (increased) demand for US assets.
Posted by: knzn at March 30, 2006 10:25 AM
knzn: Thanks for the insightful comments. I have two responses. First, regarding J-curve effects, I know this is a textbook staple, and I used to teach it in class. However, the lack of econometric evidence in support of such an effect, at least for the US, leads me to doubt the relevance of the phenomenon for discussing global rebalancing (See Rose and Yellen, "Is there a J-curve?" Journal of Monetary Economics, 1989; Bahmani-Oskooee and Brooks (Weltwirtschaftliches Archiv/Review of World Economics, 1999. In addition, using the data used in my 2005 paper, I don't find a difference in response to changes in the US trade-weighted exchange rate at short and longer lags, which is what is necessary for the J-curve phenomenon to exist).
Second, Plaza incorporated an implicit quid pro quo that the U.S. would get its fiscal house in order -- thereby decreasing the expected future stock of US government debt outstanding (and indeed the full employment budget surplus improved in those subsequent years, according to the CBO). The correct assumption -- given the current logjam in fiscal policy -- is an unchanged trajectory for government debt outstanding.
Posted by: menzie chinn at March 30, 2006 11:19 AM