精确的错误

人民币并非如同人们普遍认为的那样被低估

十年以来，中国的货币——人民币，一直保持以8.28元兑换1美元的汇率盯住美元。相当多的人特别是一些美国政客总在叫嚣人民币被极大地低估，所以中国的出口在国际市场上具有“不公平”的价格优势。
今年早些时候，一个威胁中国如果不同意让人民币升值，就对所有中国出口至美国的商品一律加征27.5%的关税的修正案被提交国会。参议院财政常设委员会在六月二十三日举行一个听证会。如果27.5%的数字听起来太精确而不可思议，那么这仅仅因为这是这个提案的支持者认为的15%到40%幅度的中间值。估算人民币的公平价格实在是一个可怕而需要技巧的任务。

一般来说，三方面的证据可以用来论证人民币是被低估的。首先，中国具有大笔的贸易赢余以及经常项目盈余。其次，人民币的贸易加权汇率已经急剧地下跌伴随着美圆始于2002的下跌。第三点，中国的外汇储备在过去几年中一直井喷式地激增。所有的这些暗示着中国当局已经使人民币的价值低于它的市场价格。但是，所有的一切都不能证明当前货币是超出弹性界限的的贬值。

在第一点上，虽然中国对美国有大笔的贸易顺差，但是由于它对其他国家同样存在大笔的逆差，所以总体来说，中国贸易赢余是很小的。无论如何，贸易不能被平衡到公平。赢余也许仅仅反应了国家储备率与投资率的不同。就第二点来说，即使人民币的贸易加权价值至从2002年之后那又怎样呢？它曾经在1994年与1998年之间显著地上涨过。最后对储备的建设也不能说明存在不公正的干涉。因为大部分这样的结果导致大量的热投机钱涌入中国去赌人民币升值。

无论怎么讨论一种货币的真正价值，首先价值必须被定义。最古老的定义是购买力平价理论（ppp）：这个理论是在长期来说，汇率应该与等于在任意两个国家购买一篮子普通商品和服务的价格。经济学人的大汉堡理论因为是比较原始地估算市场汇率变化而与购买力平价理论有所不同。我们最新的指数显示一个汉堡在中国比在美国便宜59%，所以人民币相对美元被低估了59%。更久经考验的ppp根据贸易货物估算出一个小点的在40%的数据。

第二个估算途径是基本要素均衡汇率理论（FEER）。这是一个与外部均衡（意味着稳定的经常项目平衡）内部均衡（意味着充分就业同时低通货膨胀率）联系在一起的汇率。计算出这个汇率要求中国的经常项目平衡，而并不是现在它表现出来的赢余。很多经济学家争辩说因为中国有相当高的投资回报，所以它应该是资本净进口国，所以它的经常项目应该是赤字体而不是现在的赢余。一个由ginie Coudert 与Cécile Couharde今年早些时候发表在法国的国际经济研究机构CEPII的研究表明：如果假定中国的稳定的经常项目赤子为gdp的1.5%，那么人民币相对于美元是被低估了44%。去年华盛顿国际经济机构Morris Goldstein的一个研究用了相似的方法提出人民币被低估15-25%.。

继续深入：那么内部均衡怎么样呢？很多涉及FEER研究中的一个难题。Morgan Stanley的一个经济学家Stephen Jen说他们假定中国是接近内部均衡。这是很难被400百万等待着挤进工业的农村失业工人认可的。即使外部均衡要求一个很大幅度的重新估值，内部均衡的标准也许部分地抵消。因为一个比较低的汇率能帮助未被开发的资源得到利用。并且这些FEER研究还存在一个缺陷。关于中国应该存在经常项目赤字的申明往往前提是假定新的资本流入会持续。但是中国一旦资本自由流动，那么为了多样化投资而前往海外投资的生意以及房产将会逆转这样的趋势。

泡沫之下

更为本质的是，jen先生并不乐意由所谓“稳定” 的经常项目赤字定义一种货币的真正价值。这在游资持续增长的今天实在是一个缺乏用处的指南。他更喜欢第三个方法，就是行为均衡汇率”法（BEER），在这套方法下，在过去经济学家将经济学上的变量都确立成汇率（当然有好几套理论基础），然后他们将这个几个流通价值的变量来估计均衡利率。

计量经济学的实验揭示决定人民币真实汇率最为重要决定性因素是中国的生产率增长与其他国家相关的预算平衡，以及净外国资产。运用这个模型，jen先生发现人民币相对美元仅仅是被低估了7%。高盛集团的经济学家运用一个相似的途径得到一个相似的结论：人民币仅仅被低估10%。

运用一系列的尺码，国际货币基金组织估计就如同beer的研究一样并没有发现强有力的证据表明人民币是被过度低估的。不确定于真实价值解释了imf与美国财政部为什么更喜欢说人民币需要变的 更“灵活” 的而不是直率地要求重新估值。长期来看汇率更灵活是符合中国的利益的，但是要求重新估值27.5%是依赖于不足信的基础。汇率的经济学不是那么单纯的。



Economics focus

Precisely wrong

Jun 23rd 2005
From The Economist print edition


China's currency may not be as cheap as is commonly believed





FOR a decade, China's currency, the yuan, has been pegged to the American dollar at a rate of 8.28 yuan per greenback. Quite a lot of people—most vocally, some American politicians—think that the yuan is enormously undervalued, so that Chinese exports have an “unfair” price advantage in global markets. Earlier this year, a bill was introduced into Congress that threatens to impose a tariff of 27.5% on Chinese goods unless the yuan is revalued by the same amount. The Senate finance committee was due to hold hearings on the currency on June 23rd. If the figure of 27.5% sounds too precise to be believed, that's because it is: it is simply the mid-point of a range of estimates of undervaluation (15-40%) of which the bill's sponsors were aware. Estimating a fair value of the yuan is a dauntingly tricky business.

Usually, three pieces of evidence are offered to support the argument that the yuan is far too cheap. First, China has large trade and current-account surpluses. Second, the yuan's trade-weighted exchange rate has declined sharply since the dollar started to drop in 2002. And third, China's foreign-exchange reserves have surged in the past couple of years.


All this may suggest that the Chinese authorities have held the value of the yuan below its market rate. None of it, though, proves that the currency is unfairly cheap. On the first point, although China runs a large trade surplus with America, its total surplus is much smaller because it runs deficits with other countries. In any case, trade does not have to be perfectly balanced to be fair; a surplus may simply reflect differences in national saving and investment rates. On the second point, so what if the yuan's trade-weighted value has fallen since 2002? It rose—and more markedly, at that—between 1994 and 1998. And last, the build-up of reserves is not proof of unfair currency intervention, because much of it is the consequence of flows into China of speculative money betting on a revaluation.

For any discussion of the “fair” value of a currency, that value first has to be defined. The oldest theory for doing this is purchasing-power parity (PPP): the idea that, in the long run, exchange rates should equalise the prices in any two countries of a common basket of tradable goods and services. The Economist's Big Mac index is a crude estimate of how far market exchange rates differ from PPP. Our latest index shows that a Big Mac costs 59% less in China than in America—ie, that the yuan is 59% undervalued against the dollar. More sophisticated estimates of PPP, based on traded goods, imply a smaller figure of around 40%.

A second approach is to estimate the fundamental equilibrium exchange rate (FEER). This is the rate consistent with both external balance (meaning a sustainable current-account balance) and internal balance (ie, full employment with low inflation). Working this out requires some idea of China's sustainable current-account balance. Many economists argue that because China has a relatively high return on capital, it should be a net importer of capital—ie, it should run a current-account deficit, not a surplus as it does now. A study by Virginie Coudert and Cécile Couharde, published earlier this year by CEPII, a French international economics institute, estimates that, if China is assumed to have a sustainable current-account deficit of 1.5% of GDP, then the yuan is 44% undervalued against the dollar. A study last year, by Morris Goldstein, of the Institute for International Economics, in Washington, DC, which uses a similar method, suggested an undervaluation of 15-25%.

Hold on: what about internal balance? A problem with many FEER studies, says Stephen Jen, an economist at Morgan Stanley, is that they assume that China is close to internal balance. That is hard to square with 400m underemployed rural workers waiting to shift into industry. Even if external balance requires a big revaluation, the internal-balance criterion may partially offset this, because a lower exchange rate would help to bring underemployed resources into use. And there is another flaw in the FEER studies. The claim that China should be running a current-account deficit assumes that net capital inflows will continue. But if China liberalised capital flows, these could be reversed as firms and households invested abroad in order to diversify their assets. The sustainable current-account balance might then be a surplus, not a deficit.



Beneath the froth
More fundamentally, Mr Jen is unhappy about defining a currency's fair value as that corresponding to a “sustainable” current-account balance. That is a less useful guide in a world of increasingly mobile capital. He prefers a third method, known as the behavioural equilibrium exchange rate (BEER). Under this method, economists establish which economic variables seem to have determined an exchange rate in the past (as well as having some theoretical basis), and then plug in the current values of those variables to estimate the equilibrium rate.

Econometric tests imply that the most important determinants of the yuan's real exchange rate have been China's productivity growth and budget balances relative to other countries', and its net foreign assets. Using this model, Mr Jen finds that the yuan is currently only 7% undervalued against the dollar. Economists at Goldman Sachs, using a similar approach, reach a similar conclusion: the yuan is 10% too cheap.

Using a range of yardsticks, the International Monetary Fund reckons, like the BEER studies, that it is hard to find strong evidence that the yuan is much undervalued. The uncertainty about fair value explains why the IMF and America's Treasury prefer to say that the yuan needs to become more “flexible” than to call for revaluation outright. It is certainly in China's long-term interest to make its exchange rate more flexible. But calls for a revaluation of 27.5% rest on flimsy foundations. The economics of exchange rates are rarely so simple.