我不是因为■■而抵制，我在做我应该做的！！
在疯狂抵制家乐福，LV之后
涌现了很多觉得自己应该所谓‘大度’的帖子
其实不然
抵制家乐福和LV不是因为■■而抵制
是因为我是中国人而抵制！
首先，■■没多少，别看那么多人闹，其实很多是利益的指使。
为什么欧洲人，美国人，和很多国家的人不喜欢中国人呢？！
原因很简单，
不是■■，不是中国人观念不同
他们讨厌中国人！骨子里的讨厌！！因为我们抢了他们的饭碗！
在国外，每周都能看见报纸上说哪里哪里关闭了多少商店，工厂裁了多少员工，因为工厂搬到了中国。
如果你是工人，你能不恨抢饭碗的么？
作为领导人，面对那么多失业者可能不选他，恨他
他能不想尽办法帮国家的人抢回饭碗么？
英美是聪明的，他们的公司尽量把工作分开，不让我们知道，比如，他们把call center的工作都给印度人做，工厂的工给中国人做， 近几年来，nike已经搬到印度了，因为他们怕中国学了他们技术。
有人知道么？Nike公司没有自己的制作工厂，只有他们找的批量签约工厂。
大家知道中国汇率现在为什么飞涨么？
因为美国和欧洲的逼迫！！
他们几年的逼迫下，中国汇率涨了。中国汇率是固定汇率，国家调控。
大家看到那么多的东西涨价了吧
我中国汇率课题快一年了，经过10几个不同经济模型论证和上百个专家的总结，得出
美国要毁掉我们部分公司的出口能力
为什么？
汇率没涨一点，就会有些公司无法出口东西，因为进口价提高，欧美买家只好转到其他国家，比如印度，马来西亚
可美国就不担心他们买的东西贵么？？
不会！！！！
你看看美国有多少外资在中国？！！
他们可以利用尽中国的廉价 劳动力，
中国出口东西涨价只能让这些外企用尽优势！！
中国汇率提高有很大危险
中国企业需要我们的扶持啊！！！！
真的很需要啊！！！！！！！！！！
我们真的有很多先例的！
日本曾经无奈的走过
知道 很臭名昭著的 协约叫 Plaza agreement么？？
1985 年，日本Japan,英国 Britain,法国 France, 德国Germany and the和美国 United States
这5个国家，在plaza 酒店秘密签下了plaza agreement, 最后，这个消息还是不胫而走，被日本国民臭骂了很久。
合约内容是日本增长汇率。
日本随之增长汇率，连接就是日本所有商品价格飞速上涨和滞留很久的经济，日本人忘不掉那段日子
美国经济复苏。
而现在的中国面临更大的压力！比日本当年还大！！
我不是空口说白话，对于plaza agreement 感兴趣的可以看我下面的的专家摘取研究！不好意思没中文的。
现在是我们要支持国货的时候啊！！！
大家说法国家乐福会改善中国的就业？！！
错！！！
短期的1年是，
可是长期是什么？？
是我们给他们一直打工！！！
如果中国国企不能创出品牌，
中国人蔣永远给法国人打工
给法国人做最累的活，得最少的利润！！
法国人拿着中国作的优秀产品说，看看我们法国人的东西多么优秀，多么大牌！！
中国已经有了技术，是该脱离他们的时候了！！
这样下去，在外企吞并中国企业的时候就快到了！
对不起，好像现在国企还是弱势
中国是发展中国家
还有很多人为了明天能不能饿死而担忧
不是我们可以用那么奢shi（对不起，有人会写shi么？）的LV的时候
为什么LV贵，因为国外的人工贵
不是包贵
中国人就更不该买对中国人来说物不所值的东西了
为什么假LV和真LV一样的情况下要买真的呢？
不是面子是什么？ 对不起，我可能说的过激，请原谅我，就事论事
我知道品牌是合法的，要守法，只是比喻
下面是一些有用的文章证明我不是乱讲，我做了很多研究的！！！
[ 编辑 DadLU 在 08-04-23 04:38 ]
人民币改涨么？真的么？ 
No. C2005002 2005-01
人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
No. C2005002 2005 年1 月11 日
2
人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
No. C2005002 2005 年1 月11 日
摘要：本文运用行为均衡汇率模型对人民币均衡实质汇率和汇率失调程度进行了实证研究，样本区间为1991 年1 季度–2004 年3 季度。本文的主要发现是：（1）从1994 年4 季度起人民币均衡实质汇率处于不断升值的状态，其背后的主要驱动力量是我国制造业劳动生产率的快速上升和经常项目盈余导致的净对外资产余额的不断增加；（2）1990 年代以来，人民币实际实质汇率在大部分时期偏离均衡实质汇率轨迹，表现为人民币汇率的失调。其中，1992 年2 季度–1994 年4 季度为人民币汇率低估时期， 1995 年1季度—1999 年2 季度为人民币汇率高估时期，而1999 年3 季度往后的时期人民币汇率重新转为明显的低估，并且低估程度有进一步扩大的趋势。本文对人民币汇率失调的原因分析表明1997 年以来事实上的钉住美元的汇率政策是造成人民币汇率失调的一个主要的宏观政策因素。因此，本研究的政策含义是，从应对人民币汇率失调的角度，一个更为灵活的人民币汇率制度将更有利于中国经济的健康发展。
关键词：人民币，均衡实质汇率，汇率失调，行为均衡汇率模型
3人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
摘要：本文运用行为均衡汇率模型对人民币均衡实质汇率和汇率失调程度进行了实证研究，样本区间为1991 年1 季度–2004 年3 季度。本文的主要发现是：（1）从1994 年4 季度起人民币均衡实质汇率处于不断升值的状态，其背后的主要驱动力量是我国制造业劳动生产率的快速上升和经常项目盈余导致的
净对外资产余额的不断增加；（2）1990 年代以来，人民币实际实质汇率在大部分时期偏离均衡实质汇率轨迹，表现为人民币汇率的失调。其中，1992 年2 季度–1994 年4 季度为人民币汇率低估时期， 1995 年1季度—1999 年2 季度为人民币汇率高估时期，而1999 年3 季度往后的时期人民币汇率重新转为明显的低估，并且低估程度有进一步扩大的趋势。本文对人民币汇率失调的原因分析表明1997 年以来事实上的钉住美元的汇率政策是造成人民币汇率失调的一个主要的宏观政策因素。因此，本研究的政策含义是，从应对人民币汇率失调的角度，一个更为灵活的人民币汇率制度将更有利于中国经济的健康发展。
关键词：人民币，均衡实质汇率，汇率失调，行为均衡汇率模型
一、引言
近年来，人民币汇率问题成为学术界、商界、和政策当局热烈讨论的话题。首先是东亚货币危机期间，在亚洲主要货币大幅贬值的背景下，国际金融市场和国内投资者都预期人民币将会贬值，经济学家们也就人民币是否应该贬值展开了激烈的争论。而近两年来，在中国经济快速增长，经常项目顺差不断扩大和外汇储备激增的背景下，国际上要求人民币升值的呼声不断，国内学术界和政策当局关于应否调整人民币汇率和改革人民币汇率制度的讨论也趋于激化。
理论研究表明，汇率失调（exchange rate misalignment），不论是汇率的高估还是低估，都会使经济付出福利和效率方面的代价。而大量的实证研究也表明，汇率失调会对经济产生持久、深远的影响。如Edwards & Savastano（1999）指出“持续的高估应被视为是货币危机的一个非常重要的先兆；汇率的持续失调通常是与经济在中长期内的低迷联系在一起的”，Cottani et al.（1990）运用24 个发展中国家的面板数据发现，人均收入、出口、净投资和农业的增长都与汇率失调有明显的负相关关系。事实上，从事后的角度，人们一般认为，在东亚货币危机期间人民币出现了一定程度的高估。这种高估对当时的中国经济产生了显著的影响，如出口，1998 年几乎没有增长，1999 的增长率也只有6.1%；对国外直接投资的影响更为明显，1998 年实际利用外资没有增长，1999 年出现了11%的负增长，2000 年的增长可以忽略不计。
由于汇率失调会对经济产生深刻影响，人民币汇率是否失调？失调程度如何？便是中国政策当局宏观经济管理中需要做出判断的重要问题，也是应否调整人民币汇率和改革人民币
4
汇率制度问题的核心。为了回答这些问题，我们需要知道人民币的均衡汇率，以作为判断人民币汇率是否失调，失调程度如何的标准。需要指出的是，在本文中我们关心的“人民币汇率”是人民币实质汇率（经由名义汇率转换的外国一般价格水平与本国一般价格水平之比），因为影响贸易和投资的是实质汇率，它是开放经济中最重要的相对价格之一。然而由于人民币的均衡实质汇率是不可观察的，需要我们使用科学的方法来测算。本文的目的是就人民币的均衡实质汇率和汇率失调程度进行测算，以回答人民币汇率是否失调，失调程度如何的问题。具体而言，我们关心的问题有两个，一是人民币均衡实质汇率行为与经济基本面变量之间有怎样的关系，二是人民币汇率失调情况究竟如何（东亚货币危机期间人民币是否高估，高估程度如何；当前人民币是否被低估，低估程度又如何）。并试图分析人民币汇率失调背后的经济原因。传统上经济学家们常常运用购买力平价（PPP）原理测算均衡实质汇率以及汇率失调程
度。购买力平价原理主张名义汇率是由国内外价格水平的相对变动来决定的，因而隐含着均衡实质汇率为常数的结论。该方法首先确定一个基期1，然后将该时期的实质汇率作为所有考察时期内实质汇率均衡值的估计，任一时期实际实质汇率与基期实质汇率的偏离便作为汇率失调程度的测度。这种基于购买力平价的测算方法虽然简单和直接，但并非评估汇率失调程度的好方法。一个原因是，购买力平价在现实中并不成立，如果说购买力平价的基础是商品的国际套利，由于一般价格水平计算中包括了不可贸易品，那么就没有理由认为购买力平价在现实中会成立从而构成均衡实质汇率测算的基础2。此外，这种方法无法说明均衡实质汇率本身的变化。现实中经济的基本面总是会发生变化的，政策当局感兴趣的焦点是基本面的变化是如何改变实际的和均衡的实质汇率，从而改变汇率失调的程度的。自Williamson（1985）提出基本均衡汇率 fundamental equilibrium exchange rate, 简称FEER）概念以来，许多旨在测算均衡实质汇率和汇率失调的正式方法被相继开发出来，这些方法的一个主要优越之处是能够说明均衡实质汇率本身的变动。在这些正式方法中，以Williamson（1994）为突出代表的一类方法（FEER 方法）将重点放在宏观平衡上，通过要求实质汇率与宏观平衡相一致来确定实质汇率应该达到的水平，并将该值定义为均衡实质汇率（称为基本均衡汇率）。这里，宏观平衡指经济处于充分就业和低通货膨胀（内部平衡）以及经常项目反映了可持续的净资本流动（外部平衡）这样一种理想状况。均衡汇率概念之所以被冠以“基本”一词是因为它抽象掉了短期经济因素，仅仅关注于中长期经济状况3。FEER 方法的主要缺陷是可操作性差，并由于涉及到大量与经常项目和资本项目有关的参数设定（将参数校准在充分就业和可持续的净资本流动水平），使得该法得到的估计结果对模型参数的设定比较敏感。此外，由于FEER 是规范意义上的均衡汇
率概念，FEER 方法代表了均衡汇率的规范经济学分析方法，一些被证明对实际实质汇率行为有影响的经济变量并没有包括在FEER 的计算框架中，因而FEER 方法计算的均衡实质汇率在实证意义上是否存在，即计算的均衡实质汇率是否反映了那些实质汇率的决定因素在中期的影响，是不明确的。与FEER 方法相对照，另一类测算均衡实质汇率和汇率失调的方法将重点放在实际实质汇率行为本身，通过对实质汇率有影响的相关经济变量来解释实际观察到的名义汇率和实质
1 该时期的经济被判断为处于宏观平衡状态。
2 已有的实证文献已充分证明：实质汇率的时间序列不是平稳的，并且当使用非常长的样本或使用面板数据而发现它是均值反转时，其调向用相对价格显示的均衡路径的速度也是非常慢的。因此难以用ppp 概念来解释实际实质汇率对其均衡水平的持久偏离。
3 Wren-Lewis（1992）将FEER 方法定义为“一种计算与中期宏观平衡相一致的实质汇率的计算法”。对于FEER 方法而言重要的是中期宏观平衡的概念，该方法实际上并没有对如何进行均衡实质汇率的计算规定统一的模型，实质汇率计算既可以通过大规模联立宏观经济计量模型进行，也可以通过一个局部均衡模型来计算。汇率的运动。这类方法故而被称为行为均衡汇率法（behavioral equilibrium exchange rate,简称BEER，参见Clark & MacDonald，1999）。BEER 法运用近年来计量经济学发展起来的协整（co-integration）技术，从统计学意义上发现实质汇率和早先文献识别出的各种中长期汇率决定因素之间的协整关系，以此作为确定均衡实质汇率和评估汇率是否失调的基础4。由于BEER 法只涉及到单一方程约化型（reduced form）模型的估计，较之FEER 方法具有可操作性强的优点。因此，近年来BEER 法被广泛应用于均衡实质汇率测算和汇率失调问题的实证研究（例如参见Baffes et al.，1999，Clostermann & Schnatz，2000，Maeso–Fernandezet al.，2002）。对于研究发展中国家的汇率失调问题而言，BEER 法具有进一步的优势：尽管在发展中国家的实证分析中存在诸如样本量小，数据质量不高以及经济结构不稳定等问题，但是BEER 法分析的结果常常能够如理论预言的那样发现实质汇率与基本面变量之间的协整关系，并且估计出的协整方程常常能够重现通过其它方法识别出的汇率失调情况
（Montiel，1999a）。关于人民币均衡实质汇率的测算，自东亚货币危机以来，也出现了一些规范的文献，如张晓朴（1999）、Zhang（2001）、林伯强（2002）和张斌（2003）等。Zhang（2001）和林伯强（2002）使用的是年度数据，为了增加样本的长度，他们都采用了1990 年以前的数据，甚至把样本的起点放在上个世纪50 年代，从而样本包含了各种各样的汇率体制。事实上，在1980 年代以前，价格在本质上只是政府计划的一个工具，并没有成为反映市场供需的信号。所以，对利用这些数据计算出来的结果，其可信性值得商榷。其次，张晓朴（1999）、Zhang（2001）和林伯强（2002）所使用的数据都是2000 年以前的，这样，他们的研究结果只能提供东亚货币危机期间人民币是否高估的信息，而对目前人民币是否低估不能给出回答。最后，他们对均衡实质汇率和汇率失调的情况以及背后的经济原因没有进行充分的分析，例如Zhang（2001）实际上只讨论了人民币汇率当前失调的情况5。与Zhang（2001）和林伯强（2002）不同，张晓朴（1999）和张斌（2003）使用季度数据进行人民币均衡实质汇率测算，这在数据质量上是一大改善。例如张斌（2003）采用1992 年1 季度—2002 年4 季度的样本数据，应用Baffes et al.（1999）的模型和方法估计了人民币的均衡实质汇率和汇率失调的程度。不过，他选择的基本面变量均为流量，所以其估计仅仅反映了流量均衡而忽略了资产存量的影响。此外，张晓朴（1999）和张斌（2003）还都将一些名义变量作为决定实质汇率的基本面变量（前者选择了货币供给量，后者选择了世界出口品价格），这种做法也值得商榷，因为正如超调模型（Dornbusch，1976)指出的，在长期，名义冲击不影响实质汇率6。而且货币供给只是一种政策工具，将其作为基本面变量是不合适的。本文的目的是：采用1991 年1 季度—2004 年3 季度更新的季度数据，运用行为均衡汇率（BEER）模型估计人民币的均衡实质汇率和汇率失调程度，并深入分析人民币汇率失调背后的经济原因。本文剩余部分的安排如 ：第二部分阐述我们估计人民币均衡实质汇率的计量模型和变量选择；第三部分是模型的检验和估计；第四部分对计量结果进行分析说明；第五部分给出总结性评论。
4 如果实质汇率与经济基本面变量是协整的，那么实质汇率在长期内就具有均值反转的性质，而协整方程的均值就可以视作是长期实质汇率的均衡值。
5 参见下一节关于当前失调的定义。
6 Obstfeld and Rogoff (1995)试图通过价格粘性证明实质汇率滞后现（real exchange ratehysteresis）的存在，从而证明货币冲击在长期对实质汇率有影响，但实证研究没有支持其结论（Rapach，2001）。
6
二、计量模型和变量选择
BEER 方法通过估计一个解释实际实质汇率行为的约化型方程来确定均衡实质汇率水平
和汇率失调程度，这种约化型方程的线形形式可表述如下：
t t t t q Z T â è å Œ Œ = + + （1）
其中t q 表示实际观察的实质汇率（用对数形式表示）， t Z 是那些对中长期实质汇率有影响的经济基本面变量（对数形式或比率）的当前值向量， t T 是那些对短期实质汇率有影响的短期和一时性变量（如泡沫，政策当局的发言等）构成的向量， , â è是约化型（reduced form）参数向量， tå是随机扰动项。这里我们可以看出BEER 方法与FEER 方法的另一个区别：FEER是一个中长期概念，而BEER 则更为宽泛，原则上BEER 方法也可用于解释实质汇率的周期性
运动。例如原则上可以将t t Z T â è Œ Œ + 作为短期均衡实质汇率的估计。不过由于这种短期均衡
实质汇率的估计中包括了一些泡沫因素和许多一时性因素的影响，准确估计短期均衡实质汇
率既不太现实也没有多大的政策意义。
Clark & MacDonald（1999）定义：
t t q Zâ Œ Œ = （2）
为当前均衡实质汇率（current equilibrium rate）,即该均衡实质汇率是利用经济基本面
变量的当前值（current values）计算出来的，并将实际的实质汇率t q 与当前均衡实质汇
率t qŒ 之间的差定义为当前失调（current misalignment）。当前失调可表述为：
t t cm q qŒ = - （3）
由于经济基本面变量的当前值本身也可能偏离其长期均衡水平，因此Clark &
MacDonald（1999）进一步定义
* *
t t q Z⌠= （4）
为长期（或持久）均衡实质汇率（permanent equilibrium rate），其中*
t Z 为经济基本面变
量的长期均衡值向量。将实际的实质汇率t q 与长期均衡实质汇率*
t q 之间的差定义为长期（或
持久）失调（permanent misalignment）7，长期失调可表述为：
*
t t pm q q = - （5）
BEER 方法在很大程度上强调的是实证意义，其经济基本面变量集的选择多少有些特定
7 Clark & MacDonald（1999）称之为总失调（total misalignment）。
7
或随意性（ ad hoc），因此BEER 方法的运用依赖于选择适当经济基本面变量集的理论指导。自从Edwards（1988）以来，已经出现了大量关于基本面变量选择的理论（见Montiel，1999a的一个综述），它们构成实际运用BEER 方法的基础。例如，Clark & MacDonald（1999）运用BEER 方法估计了德国马克、日元和美元实质有效汇率方程，基于Faruqee（1995）和MacDonald（1997）的研究，Clark & MacDonald（1999）认为就其研究目的而言，实质利差、贸易条件、非贸易品与贸易品的相对价格比、净对外资产以及本国与外国的政府债务比足以构成影响实质汇率的基本面变量集。有关基本面变量选择的理论成果与检验经济变量之间是否存在均衡关系的协整技术两者的结合保证了BEER 方法的有效性。在选取决定人民币均衡实质汇率的经济基本面变量时，我们考虑以下三个因素：一是理论模型所建议的变量，主要根据BEER 方法的已有文献所给出的建议（如Faruqee（1995）、Clark & MacDonald（1998）和Montiel（1999b）等）；二是数据的可得性；三是我国的具体国情。综合考虑以上三点，本文选取的经济基本面变量有：贸易条件（TOT）、非贸易品与
贸易品的相对价格比（TNT）、净对外资产（NFA）和反映贸易政策的变量（TRADE）8。下面
我们对各变量的意义及数据来源作一简单说明。
人民币实质汇率：人民币实质汇率用实质有效汇率（real effective exchange rate, 简
称REER）来衡量。用自然对数的形式表示即为：
7
1
ln( ) ln( / ) i i i
i
reer REER w S CPI CPI
=
= =‡”
其中CPI , i CPI 分别指我国和各贸易伙伴的消费者价格指数，我们考虑的贸易伙伴依次为
美国、日本、德国、英国、韩国、香港和台湾地区，它们与我国的贸易额占我国总进出口的
比重平均高达65%左右，在1997 年之前更是达到了70%。i S 表示我国与i 国（或地区）的名
义双边汇率。注意，我们这里采用间接标价法，即人民币的外币价格，名义、实质有效汇率
的上升（下降）意味着人民币的升值（贬值）。由于在1994 年以前我国实行的是双轨的汇率
制度（官方牌价与调剂市场价格共存），但外汇调剂市场承担了80－85%左右的交易量，所
以本文对1991 年1 季度－1993 年4 季度期间的名义汇率数据使用外汇交易市场的人民币价
格。i w i ＝ 贸易伙伴与我国的贸易额/7 个贸易伙伴与我国的贸易额总额。
贸易条件（TOT）：贸易条件被定义为出口价格与进口价格之比，它被用来描述一个国家
的贸易品在国际市场上的竞争力9。这里我们采用相对有效贸易条件指标，其定义为我国与
“世界”的贸易条件之比，后者通过主要贸易伙伴的贸易条件几何加权平均得到，用对数的
形式表示为：
7
1
ln ln( / ) ln( / ) i i i
i
tot TOT EX IM w EX IM
=
= = -‡”
8 这些变量都是研究均衡实质汇率问题的经济学家们一致认同的基本面变量（参见Williamson 为其1994 年
编辑著作所写的引言）。没有考虑国内外的实质利率差的原因是：到目前为止我国对资本的国际流动实行了严格管制，实质利率的影响不太显著，此外，经济学家们对是否将利率作为基本面变量尚有争议。
9 绝大多数的文献都不加说明的认为，一国贸易条件的改善（恶化）会导致该国货币的升值（贬值），但需要指出，贸易条件的改善有两个效果：一是“收入效应”，出口品价格的相对上升意味着实际收入的增加，从而更多地需求非贸易品；一是“替代效应”，进口品价格的相对下降，会增加对进口品（包括中间品和最终产品）的需求。前者推动了非贸易品价格的上升，从而有助于国内价格的上涨，而后者则有利于国内价格的下降。因此，贸易条件的改善对国内价格水平的影响是不定的，从而对实质汇率的影响也是不定的。
8
其中, EX IM分别指出口、进口价格指数。
非贸易品与贸易品的相对价格比（TNT）：该指标是一个衡量本国与外国生产率增长差
异的较为间接的指标，实证分析中也有用实质GDP 与全部劳动人口的相对比率这一更直接的指标来衡量本国与外国生产率增长差异的。直接指标试图抓住生产率增长的趋势，而间接指标则试图抓住贸易品与非贸易品部门生产率增长的差异从而体现Balassa-Samuelson 效应10。间接指标被广泛地运用于均衡实质汇率的实证研究中（如Chinn（1999），Clark &
MacDonald（1999），Clostermann & Schnatz（2000））。理论上，该指标需要使用非贸易品与贸易品的价格指数去计算。但是，在实际操作时，我们无法得到这两个价格指数。遵从文献的一般做法，使用生产者价格指数（PPI）或批发价格指数（WPI）来描述贸易品的价格变化，使用消费者价格指数（CPI）来描述非贸易品的价格走势。具体地，利用主要贸易伙伴的CPI 与PPI（或WPI）的比率，根据各自所占的权重，采用几何加权平均计算出“世界”
的非贸易品与贸易品的相对价格，然后用这一比率去除我国的CPI 与WPI 的比率，用对数表示为：
7
1
ln( ) ln( / ) ln( / ) i i i
i
tnt TNT CPI WPI w CPI WPI
=
= = -‡”
净对外资产（NFA）：把净对外资产作为实质汇率的一个决定因素，在开放经济宏观经济
学中有长久的历史，它是基于实质汇率决定的国际收支模型的考虑11。同样，在实际中，我
们得不到净对外资产的时间序列。作为替代，遵从文献的作法，利用积累的经常帐户
（accumulated current account）占GDP 的比率来模拟净对外资产的路径。然而，在我国
没有经常帐户余额的季度数据，本文的作法是用1990 年底外汇储备的余额作为1991 年初净
对外资产的替代，此后，利用每个季度的贸易盈余（出口减进口）作累计加法，得到累积的经常帐户的替代序列，最后使用这一序列与GDP 的比率，得到净对外资产NFA 的模拟数据,记这一比率为nfa。
贸易政策（TRADE）：贸易政策是影响实质汇率的一个重要因素，有大量文献发现发展中
国家贸易自由化的过程伴随着本国货币的贬值12。1990 年代以来我国贸易自由化的步伐不断加快，外贸体制改革，经常项目可兑换以及加入WTO 使我国在关税降低和非关税壁垒撤废等方面取得了实质进展，这些贸易政策的变化必然会对人民币实质汇率产生影响。然而很难到一个变量能够全面、有效地模拟贸易政策的影响。文献通常的做法是，利用进出口总额占GDP 的比率来描述一个国家的开放政策，并用该比例（即开放度）来模拟贸易政策对汇率的影响（如参见Elbadawi(1994)和Zhang（2001））。因为给定其它条件，越是自由化的贸易体制，贸易量也越大。在本文中，我们遵从这一做法。记这一比率为trade。
样本区间为1991 年1 季度至2004 年3 季度，数据为季度数据。所有的价格指数（包括
10 该效应是指贸易品与非贸易品部门生产率增长速度的差异会导致实质汇率的变化。其出发点是假设一国
的技术进步集中在贸易品部门，而非贸易品部门的技术创新则相对滞后。当前者的生产率提高时，其劳动
的边际产品增多，导致工资上升。如假设劳动力在部门之间自由流动，贸易品部门的工资上升会导致非贸
易品部门工资的上升，这样，整个非贸易品部门的成本上升，其结果是非贸易品价格的上升。而非贸易品
价格的上升又导致了国内总体价格水平的上升，从而带来本币的实质升值。
11 连续的经常项目赤字会导致该国净对外负债的增加，这需要用将来的贸易盈余来偿还。而实质汇率的贬
值将有助于产生这种贸易盈余，这意味着净对外资产余额的恶化导致中长期实质汇率的贬值。相反，本国
净对外资产余额的增加，会促进本国货币中长期的实质升值。基于跨时交易的实质汇率决定模型也将净对
外资产余额作为决定均衡实质汇率的重要基本面变量（如Obstfeld and Rogoff，1996）。
12 一般而言，封闭的落后国家能够动用的外汇储备非常有限，为了购买它们所急需的先进技术和关键设备，
不得不人为的制定一个较高的汇率，并且实行严格的贸易限制，以压制国内对国外普通商品的进口需求。
但在贸易自由化开始以后，这种过高的汇率便不再能够维系下去——外部平衡要求本国货币贬值到一个较
低的水平。
9
汇率）以1991 年1 季度为100，之后以此为基计算。国外及香港台湾地区的相关数据来源
于其政府官方统计网站； 我国的相关数据来源于高校财经数据库网站
（www.bjinfobank.com），贸易条件的季度数据来源于宋国青教授。图1-5 给出了上述变量
的时间序列图形。
图1 图2
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
1992 1994 1996 1998 2000 2002 2004
reer
4.52
4.56
4.60
4.64
4.68
4.72
1992 1994 1996 1998 2000 2002 200
tnt
图5
.2
.3
.4
.5
.6
.7
.8
1992 1994 1996 1998 2000 2002 2004
trade
图3 图4
.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1992 1994 1996 1998 2000 2002 2004
nfa
4.40
4.45
4.50
4.55
4.60
4.65
4.70
4.75
1992 1994 1996 1998 2000 2002 200
tot
10
三、检验和估计13
在运用协整方法来确定实质汇率与经济基本面变量之间的长期均衡关系之前，需要先对
各个时间序列的稳态性进行分析，即运用单位根检验来判断数据的随机性质，然后建立一个
向量自回归模型（vector autoregression, 简称为VAR），最后利用协整方法确定各变量在
长期内的均衡表达式。
（一）、单位根检验
我们利用Enders（1995）给出的步骤对各变量进行单位根检验，选择适当的滞后阶数使
回归残差尽可能的接近白噪声。结果如表-1：
表-1 单位根检验（1991：q1－2004：q3）
截距时间趋势滞后阶数ADF 统计量10%关键值
reer 有无4 -2.231161 -2.5997
nfa 有有4 -3.076352 -3.1828
tnt 有有2 -2.256945 -3.1804
tot 有无0 -2.469564 -2.5970
trade 有有4 -1.095760 -3.1828
截距时间趋势滞后阶数ADF 统计量1%或5%关键值
D（reer） 无无4 -2.557282 -1.9480
D（nfa） 无无1 -2.537237 -1.9474
D（tnt） 无无1 -3.464298 -2.6100*
D（tot） 无无3 -3.343794 -2.6120*
D（trade） 无无3 -2.658581 -2.6120*
* 表示在1%的显著性水平下拒绝单位根假设；D 表示一阶差分。
从表-1 可以看出，各变量序列都存在单位根，而它们的一阶差分都在1%的显著水平下
拒绝了单位根假设，从而各变量都是I（1）序列。这为后面的协整检验提供了良好的基础。
（二）、VAR（2）估计
协整检验的前一步工作是估计一个不受限制的VAR （ unrestricted vector
autoregression）模型。鉴于样本长度的限制，本文选滞后阶数p=2，即估计VAR（2）。虽
然使用的是季度数据，p 取4 并不能带来更好的结果，相反，模型拟合的精度在下降（其主
要原因可能是样本点的限制）。考虑到季节性的影响，在模型中加入了截距项和三个集中化
的季节虚拟变量（censored seasonal dummies）14。表-2 给出了对VAR（2）系统的诊断结
果。
13 我们运用的计量软件是EViews 4.0。
14 如果加入的是0-1 型虚拟变量，则会改变VAR 中序列的均值和趋势；如引进集中化的虚拟变量，则只会
改变均值，对序列的趋势没有影响，详见Johansen（1995）。
11
表-2 VAR（2）模型的评估诊断（Model Evaluation Diagnostics）
Multivariate Diagnostic Test
Autocorrelation LM Test LM（1） LM（4） LM（8） LM（12）
36.23631 36.29299 33.58456 23.02931
（0.0681）* （0.0673） （0.1171） （0.5759）
Heteroskedasticity Test 2 (345) ÷ =377.1188 （0.1128）
Jarque-Bera normal Test 2 (10) ÷ =17.12299 （0.0717）
* 括号里的值表示p 值。
由表-2 所给出的评估诊断不难发现：异方差检验、8 阶和12 阶自相关检验都能通过，
但1 阶自相关检验以及正态性检验的p 值较小，表明残差序列出现1 阶自相关和非正态性的
可能性较大，这种风险可能会给我们后续的估计结果带来负面影响。通过分析数据我们发现，
实质有效汇率在1993 年第2 季度降到了最低点，并且在这一时点的前后分别出现了迅速的
下降和上升的过程。而在1998 年第1 季度，人民币的实质有效汇率上升到历史的最高水平，
之后开始下降。这表明可能出现了结构性的变化（structural breaks），如在我国汇率并轨
的前夕和东亚货币危机发生的时期，汇率的时间序列数据出现了结构性的变迁。为了反映这
一个问题，我们在上述VAR（2）模型中的1993 第2 季度和1998 年第1 季度加入两个0-1
型的虚拟变量，并重新估计。结果表明VAR（2）模型改进了很多。只有1 阶的自回归检验
的p 值较低，但在通常的显著性水平下能够拒绝1 阶自相关的假设。其它的自回归检验、White
异方差检验及Jarque-Bera 正态性检验都能较好地满足进一步计量分析的需要（参见文末的
附表）。不过根据两个模型得出的协整方程系数及其标准差的估计都非常接近，计算出的均
衡实质有效汇率几乎完全一样，所以我们下面仅报告对上述VAR（2）模型进行协整检验和
估计的结果。
（三）、协整检验及其经济解释
本文采用Johansen 最大似然估计法进行协整检验，为此首先需要确定协整的具体形式15，
如截距项是否限制在协整空间里，协整变量是否采用具有趋势的形式。本文在比较了几种结
果之后，确定了最终的检验形式：协整变量具有线性趋势并且截距项限制在协整空间里。表
-3 给出了对上述VAR（2）模型的检验结果。从该结果可以看出，无论是迹统计量，还是最
大特征值统计量，都表明存在着一个协整关系。
表-3 VAR（2）的协整检验结果
协整秩H0 迹统计量5%临界值1%临界值
r=0** 78.44699 68.52 76.07
r
China’s exchange rate debate **
James Laurenceson* & Fengming Qin, China’s exchange rate debate. East Asia Economic Research
Group† Discussion Paper No. 1, August 2005 (revised September 2005), School of Economics, The
University of Queensland. Queensland.
Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.
Abstract
This paper reviews and critically comments on the policy debate surrounding China’s
exchange rate regime. There are presently two key issues – firstly, whether the RMB has
become significantly undervalued, and secondly, whether China would benefit from
adopting a flexible exchange rate regime. We find little rigorous evidence in support of
the first proposition. With respect to the second, the consensus view is that a flexible
exchange rate is desirable as it would support macroeconomic stability by providing
greater monetary independence. Most absent in this position is convincing evidence that
exchange rate stability is associated with macroeconomic instability, either in the context
of China or broader international experience. The consensus position also appears to
understate many of the benefits that accrue to China as a result of having a stable
exchange rate, as well as many of the costs associated with moving to greater flexibility.
We conclude that while adopting a flexible exchange rate regime may pass the
cost/benefit test sometime in the future, for now the focus ought to firmly be on domestic
financial reform.
Key words – exchange rate, China
JEL Codes – E58, F31
James Laurenceson *
The School of Economics
The University of Queensland
Brisbane Queensland 4072
Australia
Ph – (+617) 3365 6085
[email protected]
Fengming Qin
School of Economics
Shandong University
China
_______________________
* Corresponding author
** This paper was originally presented at the 17th annual conference of the Association for Chinese
Economic Studies (Australia) held at the University of Western Australia, Perth, July 7-8th 2005. The
authors are particularly grateful for helpful comments received from Guonan Ma and Michael Thorpe. The
paper is currently in press in Wu, Y. (ed.). 2006. Economic transition, growth and globalization in China,
Edward Elgar Publishing, Cheltenham.
† The East Asia Economic Research Group was established in July 2005, providing a focal point for East
Asia-related research of an economic nature, conducted by academic staff of the School of Economics at
The University of Queensland, their research collaborators and other interested contributors.
1
1. INTRODUCTION
One might imagine that an exchange rate left unchanged for 11 years would not generate
much interest. Yet toward the end of 2004 The Economist magazine (01/10/2004)
observed that issues surrounding China’s fixed exchange rate, in which the Renminbi
(RMB) had been pegged to the U.S dollar at a rate of RMB8.28:$US1 since 1994, had
become amongst the hottest topics in international finance. Since late 2000, much of the
interest has been prompted by speculation that China would revalue its currency. The
chief impetus for this speculation has been an accusation emanating from U.S
government circles that the dollar peg is a prominent cause of the U.S trade deficit with
China. The contention is that the RMB is pegged at an undervalued rate and is a source of
unfair advantage for Chinese exporters. In a report to Congress in May 2005, Secretary of
the U.S Treasury John Snow described China’s exchange rate policies as being “highly
distortionary” and, if left unaltered, would lead to China being labeled an exchange
manipulator under the Omnibus Trade and Competitiveness Act of 1988. Secretary Snow
stated in the report that the U.S government was calling on China to adopt a more flexible
exchange rate regime. European Union (EU) officials and the G-7 group have echoed this
call as the Euro in particular is seen as having been forced to bear the brunt of the dollar’s
depreciation since 2002. Between 2002 and 2004, the EU trade deficit with China more
than doubled, compared with the U.S trade deficit with China which increased by a little
over one half (WSJ, 17/05/2005). In the first half of 2005 trade disputes in textiles
became particularly prominent, with both the U.S and E.U erecting punitive measures in
May to slow a surge in Chinese imports that resulted from the phasing out of global
textiles quotas at the end of 2004.
Outside of government circles, calls for greater exchange rate flexibility also became the
norm from economists in the international organizations such as the IMF (Rajan and
Subramanian, 2004; Prasad, et al., 2005) and the Asian Development Bank (ADB) (IHT,
28/05/2005), as well as from those based in central banks, research institutes and
academia such as Roberts and Tyers (2003), Bergsten (2003), Eichengreen (2004),
Goldstein and Lardy (2004), Bernanke (2005), Roubini and Setser (2005) and Frankel
(2005). While most of these authors concur with the U.S government position that the
2
RMB is, if anything, undervalued, this is not their primarily focus. Rather, the
commonality they share is the viewpoint that greater exchange rate flexibility would be in
China’s own best interests as it would support macroeconomic stability by providing
greater monetary independence. This became a topical issue during 2003 and 2004 when
inflows of hot money betting on an RMB appreciation correlated with rapid growth in the
domestic money supply and in real estate prices in cities such as Shanghai.
In response to accusations of undervaluation, high-level officials in China initially
retorted that greater pressure from abroad for more rapid reform would only slow the
process down (e.g., People’s Daily, 13/05/2005). July 2005 saw a concession of sorts
with the RMB revalued by 2.1 percent to RMB8.11:$US1 and the announcement that the
currency’s value would be linked to an undisclosed basket of currencies. Following this
move, officials speaking in the government-run media began referring to the country’s
new “flexible” or “floating” exchange rate regime (e.g., People’s Daily, 22/09/2005). In
reality however, any newfound flexibility is limited. At the time of writing in September
2005, the most the RMB had been allowed to appreciate was RMB8.08:$US1, or less
than an additional 0.5% over the initial revaluation. Such changes amount to tinkering
around the edges and nothing like the degree of flexibility being called for by the
consensus position.
This paper reviews and critically comments on the policy debate surrounding China’s
exchange rate regime. There are two key issues – firstly, whether the RMB has become
significantly undervalued, and secondly, whether China would benefit from adopting a
flexible exchange rate regime. Section two finds that the usual justifications given for
claims that the RMB is undervalued have a poor basis in evidence. In section three we
offer a critique of the consensus position that China would now be best served by
adopting a more flexible exchange rate regime. Our primary purpose here is to present
the other side of the flexibility debate, which hitherto has been marginalized in the
existing literature. Moving away from a stable exchange rate would be an abandonment
of a policy that seemingly has served the country well for more than a decade and in our
3
estimation the case for doing so is far less convincing than the consensus position
portrays. Section four summarizes the discussion.
2. THE VALUATION DEBATE
There are four common arguments presented in support of the view that the RMB is
significantly undervalued. These include –
1. China’s large and growing trade surplus with the U.S proves that the RMB is
undervalued and that China is unfairly benefiting from trade.
The problem with this argument is that economic theory does not suggest that any
country will or should have balanced trade with each of its trading partners. This will be
dynamically determined by many factors, principally comparative advantage
considerations. Part of the increase in the U.S trade deficit with China simply reflects
comparative advantage considerations being allowed to run their course after having been
suppressed in the past by, for example, barriers to trade such as the Multifibre
Agreement. China is also a relatively new member of the global economy and its exports
are growing from a very small base. Another part of the rising U.S trade deficit with
China is the result of foreign direct investment emerging as a means of recycling a
country’s comparative advantage. U.S firms, as well as those of U.S trading partners such
as Japan and the Asian tigers (Hong Kong, Korea, Singapore, Taiwan), have all been
active in relocating labour-intensive manufacturing production to China in a bid to
remain competitive. Japan and the tigers have, in effect, transferred part of their trade
surplus with respect to the U.S to China. Quite staggeringly, in 2004 foreign-invested
enterprises in China accounted for 57 percent of the country’s total exports, up from just
15 percent in 1990. As a result, it should come as no surprise that China’s rapidly rising
share of world merchandise exports is matched almost entirely by a decline in the share
of Japan, and to a lesser extent, the tigers. According to WTO statistics, between 1993
and 2003 China’s share rose from 2.8 percent to 5.8 percent while Japan’s share and that
4
of the tigers fell from 9.6 percent to 6.3 percent and 10 percent to 9.5 percent
respectively.
China’s export performance over the past couple of decades is far less dramatic than that
of Japan’s in the post-WW2 period. Between 1983-2003, China’s share of world exports
grew by 3.5 percent. Yet between 1953 and 1973, Japan’s share grew by 4.9 percent, in
spite of having a weaker comparative advantage in labour-intensive manufactured goods
and foreign investment contributing virtually nothing to its exports. China’s overall trade
surplus is also not particularly large at around 2 percent of GDP in 2004. This reflects the
fact that while China may have a large trade surplus vis-à-vis the U.S, it has a deficit with
respect to other countries. China’s trade surplus is routinely less than that recorded by
leading OECD trading nations such as Germany and Japan. In 2003, Germany, for
example, had a trade surplus equal to 6.3 percent of GDP. Viewed in this broader
perspective, the view that China is pursuing a merchantilist development strategy (e.g.,
Kelly, 2005) looks decidedly shaky.
Irrespective of the source of the U.S trade deficit with China, it is patently clear that a
revaluation of the RMB would do little to reduce the U.S trade deficit overall, which in
2004 was in the order of $US600 billion, or 5.5 percent of GDP. China accounts for only
around 10 percent of U.S total trade (and only 3 percent of E.U total trade). As a result, a
revaluation of more drastic proportions than even the most ardent China critics are calling
for - say to the tune of 50 percent - would only reduce the dollar’s effective (i.e., trade
weighted) value by 5 percent. Yet between March 2002 and March 2005, the dollar’s real
effective value fell by 27.6 percent, a time period during which the U.S trade deficit only
widened. The answer to the problem of the trade deficit lies elsewhere, notably in raising
private and public savings rates in the U.S. McKinnon (2004, p.330) makes the selfevident
but important point that as long as the U.S household savings rate remains
unusually low and the U.S government runs a large budget deficit (3.5 percent of GDP in
2004), “…the relatively high-savings East Asian countries are virtually forced to run
export surpluses in order to lend their “surplus” savings to the United States - whatever
the exchange rate regime”.
5
It is sometimes said that China adopting a more flexible exchange rate would have a
broader impact because it would solve a coordination problem faced by other East Asian
countries. This line of thinking argues that other East Asian countries are resistant to
allowing their currencies to become more flexible (and presumably appreciate) without
China doing likewise for fear that their exporters would be undercut. There are numerous
problems with this argument however. For one, the numbers remain small. World Trade
Organization (WTO) statistics show that trade with China plus the six East Asian Traders
(Hong Kong, Korea, Malaysia, Singapore, Taiwan and Thailand) still only amounts to a
little over 20 percent of U.S total trade. Thus, a general appreciation of these country’s
currencies to the tune of 25 percent would only reduce the dollar’s effective value by
around 5 percent. Secondly, China’s export structure means that it does not heavily
compete in third-country markets with many of the East Asian countries that more or less
fix their currencies to the dollar anyway. Thirdly, this logic assumes that a coordination
failure has been behind the reluctance of East Asian countries to adopt more flexible
exchange rate regimes in the past. But the penchant of East Asian countries for
maintaining stable exchange rates is more readily explained by the fact that their mutual
development has been well-served by them. Since the 1980s Japan has been the exception
in East Asia in terms of having a genuinely flexible exchange rate and the performance of
its economy since this time has hardly been confidence inspiring for its neighbors.
Another relevant issue here is that the dollar value of overall trade flows is a poor guide
to the size and distribution of benefits. U.S consumers clearly benefit from cheap Chinese
imports and Andy Xie from Morgan Stanley has also estimated that for each dollar of
China trade the U.S value-added is six to eight times China’s. Thus, while in 2004 the
dollar value of U.S exports to China may only have been 17.7 percent the dollar value of
imports from China, the profits accruing to U.S firms are likely to have been in excess of
those accruing to their Chinese counterparts. It should also not be forgotten that more
than half of China’s exports originate from foreign-invested companies, including those
established with U.S capital. It is for these reasons that industry bodies in the U.S have
not been particularly vocal in supporting the government’s call for an RMB appreciation
6
and they have certainly been much quieter than in the Japan-bashing episodes of the early
1980s.
Finally, it is worthwhile elaborating upon the unusual way in which the statistics collated
by the U.S Department of Commerce deal with Hong Kong’s entrepôt trade. The U.SChina
Business Council notes that these statistics count the full value of Chinese reexports
from Hong Kong as being Chinese exports, despite the fact that services (simple
processing, packaging, marketing, etc) provided in Hong Kong add roughly 25 percent to
the value of the goods originally exported from China. Meanwhile, all U.S goods
exported to Hong Kong are counted as exports to Hong Kong, even those that are reexported
to China. According to Nicholas Lardy from the Institute of International
Economics, after accounting for Hong Kong’s entrepôt trade, the actual U.S trade deficit
with China in 2003 was 11.5 percent less than that recorded by the Department of
Commerce.
2. The decline in China’s real effective exchange rate since late 2001 means the RMB
must now be undervalued.
China’s real effective exchange rate fell by 14 percent between July 2001 and January
2005. A longer-term perspective however shows that this alone does not necessarily
imply the RMB is undervalued. The value of the RMB in January 2005 was the same as
in early 1996. Moreover, this level was only about 8 percent less than at the height of the
Asian financial crisis in the second half of 1997. At this time speculators were betting on
an RMB devaluation as the prevailing wisdom was that the Chinese currency had been
rendered decidedly overvalued. Debates over China’s equilibrium exchange rate over the
past decade have amply illustrated the limitations of estimates provided by economists.
Estimates of undervaluation currently range between 0-50 percent. Estimates of
overvaluation during the Asian financial crisis were similarly vague.
7
3. Productivity improvements associated with China’s economic transformation mean
that the RMB must now be undervalued.
Ceteris paribus, if over the past decade productivity had grown more rapidly in China
than in the U.S, then there would be a case for RMB appreciation. The problem though is
that it is not at all clear that this is what has happened. It is true that in the late 1970s and
1980s China was able to elicit rapid improvements in total factor productivity by
liberalizing its agricultural and non-state sectors. A study by IMF economists (Hu and
Khan, 1998) estimated that the average annual rate of productivity growth in China over
the period 1979-1994 was 3.9 percent. This compared with around 2 percent in other
Asian tigers (during 1966-1991) and 0.4 percent in the U.S (during 1960-1989).
However, Sachs and Woo (1997) warned some time ago that such simple sources of
productivity growth associated with China’s transitional economy were likely to soon be
exhausted and continued gains would be dependent upon reforming the more challenging
state-owned sector. Reforming the state sector has been the policy focus since the mid-
1990s and while progress has been made, the pace has been more gradual. Anecdotally,
the fact that higher economic growth rates over the past decade have required ever-larger
shares of GDP be devoted to investment is hardly evocative of an economy experiencing
rapid improvements in total factor productivity. Table 1 shows that the incremental
capital-output ratio in China has remained roughly constant since 1996. U.S productivity
meanwhile picked up during the 1990s. China’s experience during the 1980s also shows
how the impact of relative productivity movements on the exchange rate can easily be
swamped by other factors. By the time a unified exchange rate was adopted in 1994 and
the official rate was allowed to converge to the rate in currency swap markets at the time
(i.e., the market rate), the RMB had depreciated from RMB1.5:$US1 at the start of the
reform period to RMB8.28:$US1, in spite of any relative productivity improvements.
8
Table 1. Selected economic data
1996 1997 1998 1999 2000 2001 2002 2003 2004
1. Real GDP growth (%) 9.6 8.8 7.8 7.1 8.0 7.5 8.3 9.3 9.5
Gross capital formation
2. (% GDP)
39.3 38.0 37.4 37.1 36.4 38.0 39.2 42.3
3. Incremental capitaloutput
ratio (ie. 1 / 2)
0.24 0.23 0.21 0.19 0.22 0.20 0.21 0.22
4. Money supply growth
(%)
25.3 20.7 14.9 14.7 12.3 15.0 19.4 19.7 14.8
5. Domestic credit
growth (%)
24.6 19.8 20.0 12.1 11.0 13.6 29.3 19.6 9.2
6. Fixed investment
growth (%)
14.8 8.8 13.9 5.1 10.3 13.0 16.9 27.7 25.8
Source – National Bureau of Statistics and the People’s Bank of China.
4. The surge in China’s foreign exchange reserves proves the RMB is being held at below
equilibrium levels to boost exports.
This argument fails to distinguish between the contribution of the trade surplus to foreign
exchange accumulation and the contribution of speculative capital inflows betting on an
RMB revaluation. Over the period 2001-2004, the current account surplus accounted for
just 34 percent of total reserve accumulation while the dominant source was capital
inflows other than FDI (Table 2). A study published by economists from the IMF (Prasad
and Wei, 2005) reported that nearly 75 percent of the change in capital flows has come
from categories of flows sensitive to market expectations on the future trend of the
RMB/$US exchange rate, rather than the underlying fundamentals. Needless to say,
speculative sentiments can quickly change.
9
Table 2. China’s foreign exchange reserves - sources of accumulation
1996 1997 1998 1999 2000 2001 2002 2003 2004
Total reserves, inc.
gold ($US billion)
108 143 150 158 169 219 295 457 1 619
Total reserve
accumulation, inc.
gold (RES AC) ($US
billion)
32 36 6 9 11 50 77 162 207
Current account
balance (CAS) ($US
billion)
7 37 31 21 21 17 35 46 70
Net FDI ($US billion) 38 42 41 37 37 37 47 47 61 2
Net non-FDI 3 ($US
billion)
-13 -43 -66 -49 -47 -4 -5 69 77
GDP ($US billion) 821 903 954 999 1079 1176 1271 1412 1593
CAS (% GDP) 0.8 4.1 3.2 2.1 1.9 1.4 2.8 3.2 3.0
RES AC (% GDP) 3.9 4.0 0.6 0.9 1.0 4.3 6.1 11.5 3.9
Source – International Monetary Fund
Notes –
1. In 2003 the Chinese government used $US45 billion from its foreign reserves to recapitalize two state
banks. As a result, the 2003 figure for total reserves is the official value plus $US 45 billion. The 2004
figure is simply the official estimate. This has been done in keeping with Roubini and Setser (2005).
2. The FDI figure for 2004 it is not a net figure. It is simply inward FDI. The source is the National Bureau
of Statistics. In previous years, outward FDI recorded in the national accounts has been very small.
3. Net non-FDI is calculated as the residual of the change in total reserve accumulation minus net FDI
minus the current account balance.
3. THE FLEXIBILITY DEBATE
A consensus position has emerged which argues that China would benefit from greater
exchange rate flexibility because the increase in monetary independence it entails would
be more conducive to maintaining macroeconomic stability. Macroeconomic stability
does need to be accorded the utmost importance given that it has underpinned all of
China’s other successes during the reform period. However, the first point to note in
response is simply that international data do not suggest that flexible exchange rate
regimes outperform fixed regimes in terms of macroeconomic stability. In fact, the data
speak convincingly to the opposite effect. In a study of this issue, IMF economists,
Ghosh, et al. (1996, p.12), concluded -
“Does the exchange rate regime matter for macroeconomic performance?
The experience of IMF member countries since the 1960s suggests that it
does. The strongest results concern inflation. Pegged exchange rates are
10
associated with significantly better inflation performance (lower inflation
and less variability), and there is at least some evidence of a causal
relationship. There is, however, an important caveat. Countries that have
frequent parity changes – while notionally maintaining a peg – are unlikely
to reap the full anti-inflationary benefits of a fixed exchange rate regime”
Thus, not only does the data point to a better inflationary outcome under a fixed regime it
also appears to lend support to China’s reluctance to undertake frequent adjustments at
the behest of fluctuations in sentiment regarding the equilibrium value of the RMB.
Much of the recent shift in orthodoxy towards flexible exchange rate regimes appears to
have been motivated by the events of the Asian financial crisis. Yet this misses the bigger
picture. The same economies that experienced a relatively short period of crisis had
earlier experienced long periods of macroeconomic stability and rapid economic growth
under a fixed exchange rate regime and returned to a stable exchange rate and strong
growth once the crisis had passed (McKinnon and Schnabl, 2004). If China was looking
for policy inspiration from its neighbors, the Japanese experience would be the one that
stands out. A more flexible yen brought on largely by pressure from the U.S in the early
1980s did nothing to promote macroeconomic stability or steel the Japanese economy
against speculative activities and it continues to languish from the bursting of the bubble
economy more than a decade ago. It would be a serious misreading of the evidence to
claim that the experience of East Asia shows that economic development is best served
by flexible exchange rate regimes.
The case for a flexible regime also appears to have been given a popularity boost by
increasingly mobile international capital flows. It is often said that given the sheer
volume of international capital flows these days if speculators feel a currency is
incorrectly valued it would be futile for a central bank to try and defend it. For Mundell
(2003), this misses the point. Credibility is the key issue. Mundell points out that we do
not see any speculative capital movements within countries as the exchange rate
domestically is entirely credible. If a peg is credible, speculation will in fact be
discouraged. Juxtaposed against countries such as Thailand during the Asian financial
11
crisis, China’s has a healthy stockpile of foreign exchange reserves and is running current
account surpluses. In the current environment, China’s credibility will only be lost if it
becomes clear that exchange rate stability is costing the economy excessively in terms of
macroeconomic stability.
This leads to our second objection to the consensus position. Despite suggestions to the
contrary, it is not at all clear that exchange rate stability has contributed to
macroeconomic instability in China in a significant way. According to the consensus
view, as hot money has flowed into China the People’s Bank of China (PBC) has been
forced to buy dollar assets to maintain the exchange rate stability and this has resulted in
a rapidly expanding domestic money supply, excessive fixed asset investment and
increases in inflationary pressure, particularly with respect to asset prices. But as HSBC
(2005) has pointed out, total foreign capital inflows in 2004 were only equal to around 20
percent of the total value of fixed asset investment. If dampening inflationary pressure
and slowing the rate of fixed asset investment is the goal, domestic credit is the most
obvious place to start. This was precisely the approach taken by the Chinese authorities in
2004 with the growth rate in domestic credit falling from 19.6 percent in 2003 to 9.2
percent in 2004 (Table 1). Inflation which began rising in 2003 and 2004 began to fall in
2005. World Bank (2005) observes that at the end of the first quarter of 2005, the growth
rate in the domestic money supply had slowed to within the target range set by the
monetary authorities. Also acting to limit the inflationary pressures wrought by foreign
capital inflows has been sterilization activities undertaken by the central bank. Stephen
Green from Standard Chartered Bank has estimated that the PBC sterilized 47.5 percent
of foreign exchange inflows during 2004 and around 70 percent in the first half of 2005.
The costs of this sterilization have also been extremely low (see Green, 2005).
The fact that exchange rate stability can help to anchor the domestic price level appears
to have been forgotten. Xu (2000) showed that a striking long run correlation exists
between movements in the domestic price level and the real exchange rate dating back to
the start of the reform period. Xu interprets this relationship to be a bi-causal one. Before
the adoption of a unified exchange rate in 1994, changes in the official exchange rate
12
followed domestic price level fluctuations (i.e., inflationary episodes forced
devaluations). Since 1994 when the official rate was allowed to depreciate to the
prevailing market rate and by which time China’s global trade linkages had strengthened,
stability in the exchange rate has helped to secure the domestic price level. Indeed, for all
the talk of inflationary pressure in the consensus literature, it is odd that few have sought
to explain why actual inflation remains low. For Mundell and McKinnon, the answer is
plain enough – the stable exchange rate is doing its job in serving as a price anchor and is
doing it very well. The price anchor role of the exchange rate is sometimes dismissed on
the basis that bilateral trade with the U.S represents only a fraction of China’s total trade.
But this misses the point made repeatedly by McKinnon that the overwhelming majority
of trade within East Asia is invoiced in $US and that other countries in the region (with
the notable exception of Japan) also more or less peg to the dollar.
Our third criticism of the consensus view is that it understates the importance of
institutions in managing a flexible exchange rate regime. On the one hand, in highlighting
the dangers posed by hot money inflows, the consensus literature frequently refers to the
difficulties faced by the PBC in undertaking effective sterilization when domestic
financial markets are underdeveloped. Yet at the same time it calls upon the PBC to use
these financial markets to target inflation through open market operations as is done in
most OECD economies. Central banks in OECD economies have at their disposal a
powerful monetary transmission channel that results from having a complete set of
financial markets where interest rates are market determined and where borrowers are
sensitive to changes in the cost of borrowing. In contrast, China’s financial markets are
shallow, incomplete, highly regulated (see Bottlier, 2003) and the major borrowers, the
state-owned enterprises, in many cases still do not face a hard budget constraint. Thus,
the success of inflation targeting in OECD economies over the past couple of decades is
of limited relevance to China today. To be sure, as Green (2005) has pointed out,
progress in China’s domestic financial markets is being made, but for good reason
monetary policy in China continues to rely heavily on direct administrative controls such
as formal (and informal) limits to credit growth rather than indirect measures such as
interest rate adjustments.
13
In China there is also a distinct absence of financial markets that perform risk
management roles such as hedging against exchange rate fluctuations. While bankers and
traders in OECD economies have ready access to instruments such as exchange rate
futures contracts that can protect them against undesirable exchange rate fluctuations, in
developing countries such as China these agents rely on the de-facto hedge of a stable
exchange rate. In the aftermath of the Asian financial crisis, using a stable exchange rate
to provide a hedge was criticised on the basis that it might worsen the moral hazard in
domestic banks and encourage them to over-borrow in foreign currency. McKinnon and
Schnabl (2004) have pointed out however that the risk premium in domestic interest rates
is dependent upon how stable the domestic currency is relative to the currency of
borrowing, i.e., the $US. As a result, if the cross rate varies erratically, domestic interest
rates will be higher and so to will the margin of temptation to over-borrow in foreign
exchange. For this reason they conclude it is not possible to say a priori whether a stable
exchange rate worsens the moral hazard in poorly regulated banks to over-borrow. In any
case, the key issue is the effectiveness of banking sector prudential regulation not the
exchange rate regime.
A final concern we have with a flexible exchange rate has to do with the implications
increased volatility would have on Hong Kong - the showcase of the one country, two
systems approach and an autonomous region that operates a hard peg to the dollar. Hong
Kong is the classic textbook example of a small, open country that benefits from a stable
exchange rate vis-à-vis its trading partners Hong Kong’s trade is more than two and a
half times the size of its GDP. According to Hong Kong trade statistics, in 2004 total
trade (direct and entrepôt) with the mainland accounted for 43.7 percent of its total,
followed by trade with the U.S at 11 percent. It is often asserted that because China’s
exports have a high imported component, an RMB appreciation would only marginally
impact on export growth. Yet given the dependence of Hong Kong on trade with the
mainland even a modest appreciation would be of concern for the much smaller, more
trade dependent economy. Hong Kong has also been by far the largest “foreign” investor
in the mainland with the Hong Kong Trade Development Council claiming that at the end
14
of 2004, 47 percent of overseas registered projects on the mainland had Hong Kong
connections. While the consensus literature tends to cite econometric studies which
suggest that on average FDI and exchange rate fluctuations are only weakly related,
authors such as Mundell (2003) and McKinnon and Schnabl (2003) prefer to point out
case studies closer to home that may well be considered more pertinent by China’s policy
makers. Japanese FDI into many other East Asian countries, for example, has tended to
closely follow trend movements in the ¥ /US$ exchange rate.
4. CONCLUSION
The exchange rate debate in China has emerged as one of the most talked about topics in
international economics. In our reading of the existing literature, there are numerous
myths purporting to be facts and the debate regarding the appropriate degree of exchange
rate flexibility is more one-sided than is desirable. There is little solid evidence that
China’s currency is undervalued and even if it were, given the variation in equilibrium
exchange rate estimates offered by economists, reluctance on the part of China’s policymakers
to significantly revalue the RMB is unsurprising. Many of the benefits currently
accruing to China as a result of a stable exchange rate also appear insufficiently
recognized (e.g., an anchor for the domestic price level) and similarly many of the costs
involved in moving to a flexible regime (e.g., institutional constraints). Perhaps the most
prominent shortcoming of the consensus position is that it fails to convincingly
demonstrate how exchange rate stability is at the root of problems in China’s economy
today (e.g., inflationary pressure). China’s economic performance over the past decade
suggests that it has not been hopelessly trying to reconcile the “irreconcilable trilemma”
from macroeconomic theory, which states that a country cannot simultaneously pursue
free capital mobility, a fixed exchange rate and an independent monetary policy. While
its capital controls are certainly porous to a degree, when combined with partial
sterilization and monetary policy in which administrative tools remain effective, China
has been able to maintain both maintain both macroeconomic stability and a stable
exchange rate.
15
China is also fortunate in the sense that its high savings rate, cheap labor force and
attractive domestic market means that it does not face the same opportunity cost other
developing countries might when retaining capital controls (Laurenceson, 2005). The
usual argument underlying the position that even developing countries are best served by
flexible exchange rates is that it will allow them to use an independent monetary policy to
maintain macroeconomic stability while removing capital controls, with the assumption
being that the benefits of integrating into global financial markets (e.g., investment
funding, consumption smoothing, etc.) more than outweigh the costs of abandoning a
stable exchange rate. But with a savings to GDP ratio consistently around 40 percent,
China already has ample savings to fund investment. The problem for the domestic
financial sector has always been one of using existing savings more efficiently rather than
the need to mobilize more. Also, it is incorrect to say that China has not liberalized
capital controls. Restrictions over FDI have been gradually liberalized to the extent that
in 2002 China received more FDI than any other country in the world. Consequently, the
opportunity cost to China of maintaining the capital controls that support exchange rate
stability is foregoing access to more non-FDI capital (that in aggregate it does not really
need anyway) and the chance for domestic savers to earn higher returns abroad. Given
that macroeconomic stability, foreign trade and FDI have underpinned the rapid growth
in living standards during the reform period, forgoing the opportunities of higher returns
abroad is likely to be considered an acceptable sacrifice by the average Chinese saver.
In the longer term, moving to a flexible exchange rate regime may pass the cost-benefit
test. Once China’s institutional environment has been bolstered, for example, a managed
float will become more appealing. Yet based on what we know about the economy at this
point in time and the lessons learned from other countries, the area most urgently in need
of policy attention is domestic financial reform - strengthening prudential regulation,
shoring up the capital base of the banks, resolving ownership ambiguities, instituting
effective corporate governance structures and building more complete, unfettered, liquid
and transparent direct financial markets.
16
REFERENCES
Bergsten, F. 2003. The correction of the dollar and foreign exchange intervention in the
currency markets. Testimony before the committee on small business, United States
House of Representatives, Washington, D.C., June 25th. Available at -
www.iie.com/publications/papers/bergsten0603-2.htm (most recently accessed
30/05/2005).
Bernanke, B. 2005. Monetary policy in a world of mobile capital. Cato Journal, 25(1), 1-
12. Available at - www.cato.org/pubs/journal/cj25n1/cj25n1-1.pdf (most recently
accessed 30/05/2005)
Bottelier, P. 2003. China’s emerging debt markets: facts and issues. A paper presented at
China Economic Policy Reform Conference, Stanford University, 18-20 September
2003. Available at – http://scid.stanford.edu/events/China2003/Bottelier.pdf (most
recently accessed 30/05/2005)
Eichengreen, B. 2004. Chinese currency controversies. CEPR Discussion Paper No.
4375. Available at – http://ssrn.com/abstract=549261 (most recently accessed
30/05/2005)
Frankel, J. 2005. On the Renminbi: the choice between adjustment under a fixed
exchange rate and adjustment under a flexible rate. NBER Working Paper No.
11274. Available at - www.nber.org/papers/w11274 (most recently accessed
30/05/2005)
Ghosh, A., Gulde, A., Ostry, J and Wolf, H. 1996. Does the exchange rate regime matter
for inflation and growth? Economic Issues paper 13, International Monetary Fund,
Washington, D.C. Available at – www.imf.org/external/pubs/ft/issues2/ (most
recently accessed 30/05/2005)
Goldstein, M. and Lardy, N. 2003. Two-stage currency reform for China. Asian Wall
Street Journal (12/09/2003). Available at -
http://www.iie.com/publications/papers/goldstein0903.htm (most recently accessed
30/05/2005)
Green, S. 2005. Get ready for China’s Greenspan. Far Eastern Economic Review,
July/August. Available at: http://www.feer.com/articles1/2005/0507/free/p039.html
(most recently accessed 22/09/2005).
Hong Kong and Shanghai Banking Corporation (HSBC). 2005. China Economic Insight
(April). Available at -
www.hsbc.com.cn/cn/common/download/aboutus/200504.pdf (most recently
accessed 30/05/2005)
Hu, Z. and Khan, M. 1997. Why is China growing so fast? Economic Issues paper 8,
International Monetary Fund, Washington, D.C. Available at -
www.imf.org/external/pubs/ft/issues8 (most recently accessed 30/05/2005)
International Herald Tribune (IHT) (28/05/2005) ADB chief calls for letting yuan rise.
Available at – http://www.iht.com/articles/2005/05/27/business/yuan.php (most
recently accessed 30/05/2005)
Kelly, P. (28/09/2005). Taking a new long view on China policy. The Australian.
Available at -
17
http://www.theaustralian.news.com.au/common/story_page/0,5744,16742518%5
E12250,00.html (most recently accessed 29/09/2005).
Laurenceson, J. 2005. “The globalisation of China’s financial sector: policies,
consequences and lessons” in C. Tisdell (ed.) Globalisation and world economic
policies, Serials Publications, Delhi, 402-420.
McKinnon, R. 2004. The East Asian dollar standard. China Economic Review, 15 (3),
325-330.
McKinnon, R. and Schnabl, G. 2003. Synchronised business cycles in East Asia and
fluctuations in the Yen / Dollar exchange rate. The World Economy, 26 (8), 1067-
1088.
McKinnon, R. and Schnabl, G. 2004. The East Asian dollar standard, fear of floating, and
original sin, Review of Development Economics, 8 (3), 331-360.
Mundell, R. 2003. Prospects for an Asian currency area. Journal of Asian Economics, 14
(1), 1-10.
People’s Daily (13/05/2005). China never yields to outside pressure on RMB exchange
rate: Premier Wen. Available at -
http://english.people.com.cn/200505/16/eng20050516_185298.html (most recently
accessed 30/05/2005)
People’s Daily (22/09/2005). RMB exchange rate features “two way fluctuation, more
flexible”. Available at -
http://english.peopledaily.com.cn/200509/22/eng20050922_210064.html (most
recently accessed 22/09/2005)
Prasad, E., and Wei, S. 2005. The Chinese approach to capital inflows: patterns and
possible explanations. IMF Working Paper (WP/05/79). Available at -
http://imf.org/external/pubs/ft/wp/2005/wp0579.pdf (most recently accessed
30/05/2005)
Prasad, E., Rumbaugh, T. and Wang, Q. 2005. Putting the cart before the horse? Capital
account liberalization and exchange rate flexibility in China. IMF Policy
Discussion Paper (PDP/05/01). Available at -
www.imf.org/external/pubs/ft/pdp/2005/pdp01.pdf (most recently accessed
30/05/2005)
Rajan, R. and Subramanian, A. 2004. “Exchange rate flexibility is in Asia’s interest”
Financial Times, 26th September. Available at -
http://www.imf.org/external/np/vc/2004/092604.htm (most recently accessed
30/05/2005)
Roberts, I. and Tyers, R. 2003. China’s exchange rate policy: the case for greater
flexibility. Asian Economic Journal, 17 (2), 155-184.
Roubini, N. and Setser, J. 2005. Will the Bretton Woos 2 Regime Unravel Soon? The risk
of a hard landing in 2005-2006. Available at -
http://ideas.repec.org/a/fip/fedfpr/y2005ifebx13.html (most recently accessed
30/05/2005)
Sachs, J. and Woo, W.T. 1997. “China’s economic growth: explanations and the tasks
ahead”, in Joint Economic Committee, Congress of the United States, China’s
Economic Future, M.E.Sharpe, New York, 70-100.
18
The Economist (01/10/2004) A fair exchange? Available at -
http://www.economist.com/finance/displayStory.cfm?Story_id=3219515 (most
recently accessed 30/05/2005)
Wall Street Journal (WSJ) (Eastern Edition) (17/05/2005) Chinese juggernaut hits
Europe head on; strong Euro fuels trade gap, adds to continent’s economic malaise.
World Bank. 2005. China Quarterly Update April 2005. Available at -
www.worldbank.org.cn/English/Content/cqu04-05-en.pdf (most recently accessed
30/05/2005)
Xu, Y. 2000. China’s exchange rate policy. China Economic Review. 11 (3), 262-277.
[ 编辑 DadLU 在 08-04-23 04:12 ]
看看日本的血泪史吧！
中国该走跟日本一样的道理么？？
IMES DISCUSSION PAPER SERIES
Can a Rapidly-Growing Export-Oriented Economy Smoothly Exit an Exchange Rate Peg?
Lessons for China from Japan’s High-Growth Era
Barry Eichengreen and Mariko Hatase
Discussion Paper No. 2005-E-9
INSTITUTE FOR MONETARY AND ECONOMIC STUDIES
BANK OF JAPAN
C.P.O BOX 203 TOKYO
100-8630 JAPAN
You can download this and other papers at the IMES Web site:
http://www.imes.boj.or.jp
Do not reprint or reproduce without permission.
NOTE: IMES Discussion Paper Series is circulated in order to stimulate discussion and comments. Views expressed in Discussion Paper Series are those of authors and do not necessarily reflect those of the Bank of Japan or the Institute for Monetary and Economic Studies.
IMES Discussion Paper Series 2005-E-9
August 2005
Can a Rapidly-Growing Export-Oriented Economy Smoothly Exit an Exchange Rate Peg? Lessons for China from Japan’s High-Growth Era
Barry Eichengreen†and Mariko Hatase‡
Abstract
We explore the parallels between Japanese currency policy after World War II and Chinese currency policy today. After two decades of pegging at 360 yen, Japan decoupled from the dollar on August 1971 and then repegged at a revalued rate of 308 yen. After stabilizing the exchange rate at this new level for about a year, greater flexibility was introduced. This phased adjustment – revaluation followed after a time by an increase in flexibility – bears more than a passing resemblance to recent Chinese policy initiatives.
We analyze the impact of Japan’s exit from its peg on exports and investment. The results point to sizeable effects of the yen’s revaluation on both variables, especially investment. While our analysis suggests that a rapidly-growing, export-oriented economy can operate a heavily managed float despite the presence of capital controls and the absence of sophisticated foreign currency forward markets, it underscores the importance of managing the exchange rate with domestic conditions in mind and avoiding the kind of large real appreciation that would sharply compress profits and damage investment.
For China this suggests starting with a modest band widening and a limited increase in flexibility, and not with a large step revaluation which could have a sharp negative impact on investment and growth. Our results thus provide support for the kind of measures taken at the end of July.
Key words: China, Japan, exchange rate, peg, exports, investment
JEL classification: F31, F33, N15, N65
† Professor, University of California, Berkeley (E-mail:eichengr@econ. Berkeley.EDU)
‡ Director and Senior Economist, Institute for Monetary and Economic Studies, Bank of Japan (E-mail:[email protected])
The views expressed here are not those of the Bank of Japan or any other organization with which the authors are affiliated. We are grateful to Miguel Fuentes and Yoichi Matsubayashi for their help with some technical aspects of the paper. For helpful comments on an earlier draft we thank Mitsuhiro Fukao, Yutaka Kosai, Akira Otani, Shigenori Shiratsuka, Miyako Suda, Wataru Takahashi, Yosuke Tsuyuguchi, participants of the seminar at the Bank of Japan and the staff of the Institute for Monetary and Economic Studies of the Bank of Japan.
Contents
1. INTRODUCTION…1
2. BACKGROUND TO THE STUDY…2
3. CURRENCY POLICY…11
4. EXITING THE PEG…16
5. CAPITAL CONTROLS AND THE FORWARD MARKET…20
6. EFFECTS OF EXITING THE PEG…24
7. CONCLUSION…33
1. Introduction
China’s exchange rate arrangement is a continuing topic of controversy, notwithstanding the decision announced on July 21st, when the country revalued by two per cent and indicated that henceforth the rate would be allowed to depend more heavily on market conditions. Is the initial revaluation too small, or would a larger adjustment have had a devastating impact on Chinese investment and exports? Is the two per cent revaluation properly seen as part of a larger process, as a first step in the transition to greater flexibility? And, if so, is this gradual approach in which both the adjustment of the exchange rate’s level and the move to greater flexibility are phased in an appropriate strategy by the Chinese authorities?
There are few historical precedents for these questions. It is hard to think of many large, export-oriented, fast-growing economies in the early stages of catch-up that exited voluntarily from a peg. One analogy is Japan in the 1970s. After two decades of pegging at 360 yen to the dollar, Japan decoupled from the dollar on August 28th, 1971, repegging on December 18th at 308 yen in conjunction with the Smithsonian Agreement. The new peg lasted 14 months, after which greater flexibility was introduced.
In this paper we use this historical precedent to shed light on current Chinese prospects. We argue that there are extensive parallels between the two cases, extending even to the gradual nature of the adjustment – a one-time step revaluation followed after a period by the shift to freer floating. At the same time, the analogy must be developed carefully. Japan then was more advanced than China now. Its per capita incomes were higher, and technological sophistication was greater relative to the United States. While government involvement in the economy was extensive, such involvement was not as
1
pervasive as in China. Japan also had more sophisticated financial markets and a better-developed monetary transmission mechanism. Moreover, its initial devaluation was larger. Still, and notwithstanding these differences, Japan’s experience from the 1950s through the 1970s can provide useful insight into China’s current options.
Our analysis of the country’s experience from the 1950s through the 1970s suggests that a rapidly-growing, export-oriented industrializing economy can exit a currency peg without significantly disrupting its growth, its export growth in particular. It can operate a more flexible exchange rate despite the maintenance of capital controls and not having well-developed foreign currency forward markets. But it is important for the country to manage its exchange rate with domestic economic conditions in mind and to avoid the kind of large real appreciation that could have a sharply negative impact on profitability and investment. For China, this suggests starting with a modest revaluation and a progressive widening of the band – the policies actually adopted – is preferable to the kind of large revaluation sometimes advocated by foreign observers.1
2. Background to the Study
Japan in the 1950s and 1960s, like China in the 1980s and 1990s, grew at rates well in excess of those witnessed anytime in its previous history.2 GNP growth in Japan between 1955 and 1971 averaged 9.3 per cent per annum. This was a dramatic acceleration from 1913-1950, when growth had averaged 2.0 per cent, and from 1885-
1 See inter alia Goldstein and Lardy (2003).
2 The detail provided here may be more than is required by some Japanese readers already familiar with the structure of the economy. However, it may be important to provide this information for readers interested in the comparison with China, especially when it comes to topics like the organization of the foreign exchange market, the monetary transmission mechanism, and the operation of capital controls.
2
1913, when it had proceeded at 2.6 per cent.3 China’s economy has expanded at nearly 10 per cent per annum since economic reform commenced in 1978, which is similarly a revolution in growth performance. (See Table 1.)
In Japan after World War II as in China today, growth was fueled by exports, facilitated by the transfer of advanced technology, sustained by investment, and supported by elastic supplies of cheap labor. Japan’s exports grew at an annual average rate of 16.9 per cent between 1955 and 1971, more than half again as fast as output.4 Exports quadrupled in volume between 1959 and 1969 and increased by a further one-third in the next two years, impressive performance even by modern Chinese standards. The other side of this coin was a doubling of Japan’s share of the imports of the rest of the world in the course of the 1960s and especially rapid growth in the share of U.S. merchandise imports accounted for by Japanese suppliers. (See Table 2.) Japan’s surplus with the U.S. was more than $1 billion in 1970, reflecting the buoyant U.S. market for the country’s exports of consumer goods in conjunction with Japan’s need to finance imports of energy and raw materials from other suppliers. Strong surpluses allowed Japan to more than double its foreign reserves between 1967 and 1970.5
By the end of the 1960s, Japanese gross national savings rates had risen to 40 per cent and the investment/GNP ratio had reached 35 per cent, levels broadly comparable to those in China today.6 (See Table 3.) Employment in agriculture declined by 3 per cent
3 The 1913-1950 figure is depressed by the exceptional events of the post-World War II period. Over the shorter period 1913-1940, growth averaged 4.0 per cent.
4 This nearly matches rates of export growth in China, which have been running on the order of 20 per cent per annum in recent years.
5 Though outstanding reserves had been constant during the high growth era, they started increasing in 1968. Reserves of the end of 1970 were $4.4 billion, up from $2.0 billion at the end of 1967.
6 The former showed about 50 per cent increase, while the latter achieved almost 90 per cent increase compared with two decades earlier. Between 1956 and 1973 real fixed nonresidential capital investment grew at an annual average rate of 16 per cent.
3
per annum both in the second half of the 1950s and first half of the 1960s and was offset by the rapid growth of employment in the modern sector, led by manufacturing, facilitating industries, construction, and to a lesser extent services.7 According to Ohkawa and Rosovsky (1973), more than 70 per cent of Japan’s full-time farm households were in a condition of underemployment after World War II. Elastic supplies of labor could thus be made available to industry without driving down agricultural output.8 Contemporaries referred to a dual wage structure in which the wages of employees of enterprises in the modern sector were at least twice the levels of those in the agricultural sector – indicative of a strong incentive for labor reallocation.9 In all these respects, the parallels with contemporary China are clear.
But there are also important differences. Japan at the beginning of its high-growth period was a relatively advanced industrial economy, reflecting the development of a modern textile industry in the 19th century and then steel and shipbuilding in the first half of the 20th.10 China, in contrast, had a dearth of modern industry when embarking on
7 Ohkawa and Rosovsky (1973), Table 5.2.
8 Ohkawa and Rosovsky (1973), p.133. That widespread underemployment was more than a transitory post-World War II phenomenon is evident in the rationale for the famous National Income Doubling Plan of 1960, which was precisely to reduce the extent of disguised unemployment. Prime Minister Hayato Ikeda announced the Income Doubling Plan immediately after he took office in 1960. The plan was crystallized as a cabinet decision in December 1960 (Yasuba and Inoki 1989, p.18). Its primary goals were a sharp improvement in living standards and the achievement of full employment. The plan intended to achieve an annual averaged growth of 7.2 per cent between fiscal year of 1961 and 1966. To reach that target, the following areas were given priority; improvement of social infrastructure and industrial structure, enhancement of trade and international economic cooperation, and stimulating technology and human resources (Kosai 1989, pp.210-2).
9 More precisely, Ohkawa and Rosovsky (1973) suggest that wages in large industrial enterprises were as much as twice wages in small and medium-size firms, while wages in the modern sector broadly defined (small and large firms alike) were twice the levels prevailing in agriculture.
10 Nakamura (1993) categorizes Japanese cotton spinning as “modern industry” from the late 1880s (Nakamura 1993, p.84). Takamura (1971) concludes that the modern cotton spinning industry was established by the late 1890s (Takamura 1971, p.209). Okazaki (1993, p.227) argues that the Japanese steel industry was able to stand on its own feet, with the help of limited protection from import competition, from the late 1920s. The shipbuilding industry reached international standards by the early 20th century (Ishii 1991, p.228). Shipbuilding recovered relatively quickly after World War II; by 1955 more than 85 per cent of its products were exported (Mitsubishi Heavy Industry, Ltd. 1967, pp.7, 22, 49-51). We exclude 4
reform in 1978. Japanese per capita GDP in 1950 expressed in 1990 international Geary-Khamis dollars was $1,926, whereas Chinese per capita GDP in 1978 was $979. Five years later the comparable figures were $2,772 and $1,265.11
The Japanese system of technology transfer was also different. Whereas China today relies on foreign direct investment for technology transfer, Japan relied on licensing and reverse engineering. The Japanese government limited inward FDI for three decades after 1950. Mason (1982) and others argue that restrictions on market access for foreign multinationals facilitated the country’s efforts to license foreign technologies.12 They conclude that the relatively advanced state of the Japanese economy, together with the lower tacit component of advanced technology compared to today, enhanced the effectiveness of this approach to technology transfer.
The structure of labor markets was different as well. In Japan in the 1950s, labor was free to move from the countryside to the cities; the freedom to select one’s occupation and place of residence were individual liberties guaranteed since the early Meiji period, the late 19th century.13 In China, in contrast, the authorities attempt to regulate the movement of the rural population to the cities in order to avoid threats to the
motor vehicles and electrical machinery from this list of industries that developed significantly before World War II; although Toyota produced trucks for military use in the 1940s, the production of passenger cars only started after the war. The Ministry of Trade and Industry imposed significant restrictions on imports in order to protect the passenger car industry in the 1950s (Toyota Motor Corporation 1987, pp.83-4, 259-60). Although there was prewar production of lighting equipment, radios and electric wires, the leading companies producing electrical equipment and electronics, such as Sony and Sanyo Electric, were only established after World War II.
11 According to estimates in Maddison (2001), Table C3-c.
12 Kosai (1989) points out that one of the reasons for restricting inward FDI was the fear that this would hamper domestic research and development.
13 In the very early stage of Meiji period, a series of reforms which enabled people to select one’s occupation and to move freely conducted; the class system separating warriors, farmers, craftsmen and merchants abolished in 1869 and the trades of agricultural field were allowed in 1872. For details, see Nakamura (1993), pp.62-63.
5
public order, using a system of official work permits or visa.14 The effectiveness of these measures is difficult to evaluate, since illegal internal migration is rife. So too are the implications for the China-Japan comparison. On the one hand, China today has even larger reserves of underemployed rural labor anxious to find work in the high-wage modern sector. On the other hand, that labor’s freedom to move is significantly less.15
In addition, the banking and financial system was stronger in Japan. Banks were not burdened by the overhang of nonperforming loans plaguing the Big Four banks in China. Their lending decisions were guided by commercial motives, although subject to window guidance (quantitative limits on the growth of lending by individual banks, described in more detail below). The lending charges of Japanese banks could be adjusted freely (even if published lending rates were relatively stable, the banks still adjusted the compulsory depositions required of their customers).16 In China, banks have
14 In addition, already in the 1950s employment in large Japanese manufacturing firms was characterized by long tenures and low involuntary separation rates. See Abegglen (1958) and Odaka (1999). Labor relations in the high growth era were well known for enterprise unionism. The unions of major firms played important roles in determining wages through negotiations with management. The emphasis of these unions on job security over wage growth is often cited as one of the reasons for low levels of wage rigidity in Japan compared to other developed countries. Shimada, Hosokawa and Seike (1982) point out that rates of wage increase were sensitive to factors affecting corporate profits in Japan while U.S. wages were less flexible. Odaka (1999) points out that the co-operative attitudes of unions worked as protection against cost inflation. For example, unions took “an extremely co-operative stance towards firms’ policies of streamlining and wage restraint” after the first oil shock (p.146). China today does not possess equally well-organized union movement or comparably stable employment relations.
15 To put it another way, one rationale for maintaining an undervalued exchange rate is to offset another distortion (rural underemployment) that makes the social returns to employment in the modern export-oriented sector much higher than the returns to employment in the rural sector. See Dooley, Folkerts-Landau and Garber (2003). Since China today has more rural underemployment, it can be argued that it should avoid further revaluation. On the other hand, this argument for sticking to an undervalued rate in order to accelerate the movement of rural labor to the modern sector makes no sense when the authorities are at the same time attempting to limit internal migration for fear of social dislocations.
16 For the details of lending rate adjustments using compulsory deposits, see Ministry of Finance (1991a, pp.267-273).
6
less freedom to adjust lending rates, and in any case such rates matter less for lenders and borrowers with soft budget constraints.17
Monetary control exhibited similarities as well as differences. In the 1950s the BOJ used the discount rate, discount-rate surcharges for excess borrowing, and window guidance to influence the growth of money and credit.18 It used changes in required reserves, which altered the banks’ capacity to lend, starting in 1959.19 From the mid-1950s it also sold short term government securities. It purchased and sold long term bonds, especially after the government resumed issuing these in 1966. It purchased private bills (prime bills or promissory notes issued by banks with prime bills attached as collateral, which were introduced in 1972).
At the same time, open market operations of the normal sort did not really exist, since there was little in the way of a liquid bond market.20 The central bank rationed its
17 Although it can be argued that the bribes paid to loan officers in order to obtain bank credit serves a similar function to compulsory deposits.
18 Discount rates applied to BOJ lending to private banks (discounts and loans on bills). From 1946 to 1963, the BOJ applied discount rates to pre-determined amounts of BOJ lending and amounts over the ceiling were subject to higher interest rates. By changing the level of the ceiling, the BOJ could affect the lending behavior of banks.
19 According to Suzuki (1987, pp.322-3), “The reserve deposit requirement system is a system under which commercial financial institutions are required to deposit in non-interest bearing accounts at the Bank of Japan amounts in certain proportions to their deposits and other liabilities (these proportions are called ‘reserve ratios’). … During periods of financial tightening, and subsequent loosening or relaxation until around 1980, reserve requirements were changed relatively frequently and thus were a powerful policy instrument for the adjustment of the liquidity position of financial institutions, together with lending policy and securities and bills operations.”
20 A study that emphasizes these features of Japanese financial markets and monetary policy is Tachi and Komiya (1960, p.288). They write:
“One of the peculiar features of the Japanese monetary system and monetary policy in recent years as compared with those of other advanced countries is the prevalence of direct control measures such as credit rationing at the discount window of the Bank of Japan, the fixing of almost all rates of interest by, or under the direction of, the monetary authorities, or the regulation of new issues of bonds virtually on a quota system. In Japan, because of the absence of an open market in such short-term securities as Treasury bills and acceptance, the Bank of Japan depends on loans as the principal means of extending central bank credit to the banking system.”
Another example of this view of the financial system can be found in Bank of Japan (1976). There one finds the passage (pp.123-125, authors’ translation):
“There are three kinds of conventional tools for monetary policy: control of discount rate, open market operations, and control of required reserve ratios. Among these three, control of the
7
discounts and loans against bills. It imposed window guidance on the large city banks in the 1950s and then on the long-term credit banks, large regional banks and trust banks in the 1960s. It also used informal tools for controlling corporate bond issuance.21 When selling long term government bonds in the late 1960s, the BOJ informally allocated them to banks in specific amounts.22 This reliance on moral suasion resembles current practice in China, where the authorities similarly attempt to influence bank lending and thereby money and credit conditions by instructing the banks to limit and adjust their lending to accommodate the objectives of monetary policy.23
discount rate has been used as the most important policy tool…Open market operations cannot be conducted with flexibility due to underdevelopment of bond market. The major method of monetary control was BOJ lending with discount rate control. So called window guidance was used as a supplementary tool. … The conventional tools make it possible to control private banks’ activities through the effect of interest rate and liquidity on their cash position. However, it has not been easy to obtain immediate effects of such policies, especially for tightening occasions, as the function of interest rate and market has not been sufficient. Therefore, the BOJ used window guidance as a supplementary tool for conventional methods.”
In his assessment of monetary policy from the late 1960s through the early 1970s, Kosai (1989, p.228) concludes that “there was a sign that monetary policy tools transformed from BOJ lending, call rate and window guidance into open market operations and interest rates.”
21 From 1949 to 1956, the BOJ influenced corporate bond issuance by determining the eligibility of bonds for use as collateral. Financial intermediaries, mainly banks, purchased about 90 per cent of corporate bonds, and those bonds were immediately passed to the BOJ as collateral for BOJ lending. When monetary conditions eased in the mid 1950s, the banks’ dependence on BOJ lending declined and the BOJ’s de facto power through this mechanism disappeared. After the pre-issue examination of eligibility was eliminated, eight major banks, including Industrial Bank of Japan and the major city banks, developed informal arrangements whereby they determined detailed conditions for corporate bond issuance in conjunction with the BOJ. Thus, the central bank continued to exercise considerable influence over the volume of issuance (Ministry of Finance 1991b, pp.570-2, Association of Public and Corporate Bond Underwriters 1980, pp.232-3, 239).
22 When the BOJ broadened the practice of purchasing bonds in 1962, it allocated those transactions 75 per cent to city banks, 18 per cent to regional banks, and 3.5 per cent to trust banks and long-term credit banks. Within those categories it assigned amounts to each institution according to the size of its balance sheet (Ministry of Finance, 1991a, pp.189-190). When the BOJ began purchasing long-term government bonds in 1967, a formal ceiling on the size of the operation was not set. But the BOJ set informal guidelines that worked as a de facto ceiling. According to Ministry of Finance (1991a, p.310, authors’ translation), the BOJ considered that “about 70 per cent of the outstanding [government bonds] held by financial institutions, excluding Norin Chukin Bank and securities companies, are the appropriate ceiling for the operation.” Following this guideline, each entity taking part in such operations was allocated an amount equal to 70 per cent of its holdings of eligible bonds.
23 Typically, the People’s Bank of China convenes a meeting to explain its monetary policy intentions and uses moral suasion to guide the commercial banks to adjust their lending policies accordingly. For example, “in the first quarter of 2004, considering overinvestment in some regions and industries, the PBC held monthly meetings to review economic and financial development and strengthened warnings for the
8
To be sure, the extent of this reliance differed across markets and over time. As for short-term money markets, the call market tended to be subject to controls though the degree of restrictions changed over time. “The experts including the authorities considered that the level of call rates were abnormally high” in the early stage of the high growth era and, thus, moral suasions by the BOJ or self-imposed restrictions by financial institutions for call rates were believed to be necessary.24 However, the coverage of these restrictions was limited and there were loopholes allowing effective rates to fluctuate to some extent. Thus, “call loan rates worked as indicators for the condition of financial market as a whole though the function was imperfect.” 25As for bond markets, yields of government-guaranteed bonds, bank debentures and corporate bonds in secondary markets could fluctuate relatively freely, while those in primary markets were kept in artificially low levels. By contrast, yields of JGBs in both primary and secondary markets were strictly controlled.26 But despite these variations, and notwithstanding the
commercial banks to guard against potential risks. Commercial banks are urged to maintain sound operations and sustained business development with intensified capital constraints and reasonable loan growth. ” In addition to monthly meetings, it announced a notice on 18 January, requiring commercial banks to take positive measures to control loans to such overheated industries as steel, aluminum and cement. Furthermore, it “summoned all commercial banks to carry out a specific analysis on current economic and financial situations, requiring a mechanism to be introduced to control loan expansion in line with their own risk-control capacities and capital adequacy status.” (People’s Bank of China, 2004a, p.51.) That said, there has been movement in the conduct of these policy in recent years, with the People’s Bank beginning to make growing use of interest rates to shape the growth of bank credit.
24 Ministry of Finance (1991a), p. 87. From 1948 to 1957, the BOJ “guided” the market participants to trade at a particular overnight interest rate in the call market. This de facto regulation on call loan interest rates was abolished in 1957 but then self-imposed controls by financial institutions started two years later. The effectiveness of the self-imposed controls was not satisfactory and the actual interest rates applied to transactions were higher than published rates. Thus, the BOJ again started to “guide” call loan brokers in 1962 and in the mid-1960s this control gradually moved towards quotation system under which call loan brokers decided quoted prices with the consultation to the BOJ (Ministry of Finance 1991a, pp.89, 111, 155-6,157-8).
25 Ministry of Finance (1991a), p.313.
26 Ministry of Finance (1991a), pp.313-4, Kure and Shima (1987), pp.56-7. In the case of the JGB, the amount of circulation in the secondary market was limited as most part of the JGBs were held by financial institutions and it was said that the MOF restricted the sales of them. The MOF “guided” the securities companies to buy JGBs from individuals with prices close to those in primary markets (Bank of Japan 1986, p.529, Kure and Shima 1987, pp.56-7).
9
fact that Japan had a sounder banking and financial system after World War II than China today, the mechanisms of monetary control were not entirely dissimilar.
Finally, compared to China today, government involvement in the economy was less. To be sure, volumes have been written about the industrial policies of the Ministry of Trade and Industry and the credit-allocation policies of the Ministry of Finance. The Japanese government sought to channel resources into industrial development, establishing the Reconstruction Finance Bank (eventually taken over by the Japan Development Bank) and the Export Bank of Japan (renamed the Export-Import Bank of Japan). These public financial institutions influenced Japanese development through the Fiscal Investment and Loan Programme, or FILP.27 The FILP was overseen by the Ministry of Finance, which collected public deposits at the Postal Savings system (and the resources of the public pension system) and passed them through to the JDB, the ExIm Bank, and other public financial institutions such as the Housing Loan Corporation.28
The commercial banks, in contrast, retained basic autonomy over their loans and investments within ceilings set by window guidance.29 Although they were subject to window guidance, there was no counterpart in 1950s and 1960s Japan of the nonperforming loan problem in contemporary China, where estimates of the extent of
27 For a more detailed description of the FILP, see Suzuki (1987), pp.273-276.
28 According to Suzuki (1987), the share of public banks, including the JDB and ExIm Bank, in the FILP fluctuated between 8 and 20 per cent from 1965 to 1980. Public finance corporations, such as Housing Loan Corporation usually accounted for the largest share. Outstanding of the FILP funds were 16.2 billion yen in the end of FY 1965, while all outstanding bank assets were 36.5 billion yen. While public financial institutions such as the JDB and ExIm Bank had assets comparable in size to those of Japan’s largest private banks, there were in a goodly number of private banks; postal saving deposits comprised no more than 20 per cent of personal savings. Roughly speaking, then, some four-fifths of financial intermediation in Japan was based on commercial motives.
29 There is some controversy over this. Authors like Mikuni and Murphy (2002) suggest that the banks came under strong moral suasion to plow their loans into capital formation in export industries in particular.
10
nonperforming loans are on the order of 40-50 per cent of GDP. This suggests that the extension of credit for non-economic reasons was less than in China today. While it is hard to put precise figures on the share of policy loans in the portfolios of Chinese state banks, it is fair to say that the banking system, which is heavily dominated by state banks, has devoted significantly more of its resources to development lending, not always with positive results for their own balance sheets and for the development of the Chinese economy.
3. Currency Policy
For nearly a quarter of a century after World War II Japanese currency policy was predicated on an exchange rate pegged to the dollar and stringent capital controls. The yen was pegged at 360 on April 25, 1949 in conjunction with the Dodge Line.30 It remained there until the collapse of the Bretton Woods System in 1971.
Exchange rate stability has been portrayed as integral to the Japanese economy’s rapid growth and to the rapid growth of its exports in particular.31 To be sure, in strong upswings, the demand for imports grew more rapidly than the supply of exports, and export supply was partially crowded out by domestic absorption. It was not possible to finance the resulting current account moved out of foreign reserves, since the government followed a strategy of limiting reserve accumulation in order to plow all available resources into fixed investment. This meant that the Bank of Japan had to tighten in
Others like Horiuchi (1984) are more skeptical that the authorities were in fact able to influence the allocation of bank lending in this way.
30 The Dodge Line was the set of economic reforms advanced by Joseph M. Dodge in the late 1940s. Dodge was the President of Detroit Bank and appointed as an advisor to the General Headquarters of the occupying forces (GHQ). For details of the Dodge Line, see Nakamura (2003) and Miwa (2003)…
[ 编辑 DadLU 在 08-04-23 04:35 ]
接上文
11
order to restrain the growth of demand, limit the deterioration of the external accounts, and defend the currency peg. This constraint on the rate of growth was referred to as the “balance of payments ceiling.”32 Fortunately, this constraint only bound at high rates of growth, at least after the beginning of the postwar period.
This observation has led observers to ask whether the yen was significantly undervalued. The consensus appears to be that the currency was overvalued at the beginning of the postwar period but that this gave way to undervaluation in the course of the high-growth years.33 Before the currency was pegged in April 1949, there had been debate over its appropriate level between Japanese officials and the Economic and Scientific Section of the Supreme Commander for the Allied Powers (SCAP). SCAP proposed a rate of 330 to the dollar, but Japanese officials preferred a lower rate in part because they anticipated that sterling and other currencies might soon be devalued against the dollar.34 The Bank of Japan (1985, p.258) observes that, at the eventual rate of 360 to the dollar, “the exchange rate was considerably overvalued given the effective price level considering black market prices. It was a challenge to adapt the Japanese economy to the newly set exchange rate.”35
Initially, tight foreign exchange controls were needed to support the currency. As Komiya (1988) puts it, “the exchange rate of 360 yen per dollar set in 1949 was significantly overvalued compared with the equilibrium rate at which the external accounts are balanced without restrictions on imports and without export enhancement
31 Odaka (1989, pp.175-6, authors’ translation) writes in a representative passage that “it should not be neglected that the fixed exchange rate of 360 yen against dollar was firmly kept until 1972 and that this was behind the export expansion.”
32 For references see Nakamura (1993), p.137 et seq. An analysis of this phenomenon is Ackley and Ishi (1976).
33 Note the similarity to recent discussions of China.
34 Ministry of Finance (1976), pp.430-1; Bank of Japan (1985), pp.255-6. 12
policies.”36 37 Kosai (1989) agrees with the proviso that overvaluation emerged only after sterling was devalued and inflation accelerated with the outbreak of the Korean War.38 These observations suggest at least some qualification of the conventional wisdom that an undervalued exchange rate was central to Japan’s postwar growth from the start.
But with the recovery of the economy following the imposition of the Dodge Line and the rapid growth of productivity in the traded goods sector in particular, what had once been seen an overvaluation became increasingly undervaluation.39 Government-led rationalization of the metals, machinery and chemicals sectors led to reductions in production costs in industries in which Japan already had some presence in export markets. In addition, at the end of the 1950s and increasingly in the 1960s, Japan’s export competitiveness was enhanced by technological progress and structural change leading to the emergence of new export products that had not been produced before in significant quantities or even at all, plastics, business machines and automobiles for example.40
Already at the end of the 1950s Shinohara (1959) used comparisons with the 1930s to argue in favor of undervaluation.41 While there is some debate over precisely when this undervaluation emerged, by the late 1960s there was a considerable amount of evidence in favor of the hypothesis.42 By this time Japan had learned how to produce many of the capital goods needed for its industrial expansion rather than importing them
35 Authors’ translation.
36 Komiya (1988), p.160 (authors’ translation).
37 Some previous studies, such as Shinohara (1959) and Fujino (1988) discuss the evaluation of the yen from the viewpoint of purchasing power parity.
38 Kosai (1989), p.265.
39 See inter alia Fujino (1988) and Kosai (1989).
40 On this see Kojima (1972).
41 Shinohara (1959), p.24.
42 For questions about the timing of this shift, see for example Fujino (1988).
13
from abroad. After 1965 the current account moved into surplus, and then from 1968 foreign reserves rose strongly.43 The balance of payments ceiling of which observers had warned no longer seemed to constrain growth even in strong expansions.44
To analyze these questions further, we have constructed nominal and real effective exchange rates for the 1950s and 1960s.45 Our series are trade-weighted averages of bilateral nominal rates for 17 leading trading partners accounting for 40 to 50 per cent of total Japanese exports.46 Figure 1 for the nominal exchange rate shows that focusing exclusively on the yen-dollar exchange rate, as is typically done in discussions of the post-World War II years, understates currency variability. It also shows a tendency for the nominal effective rate to appreciate over the high-growth period as a result of devaluations against the dollar and thus the yen by other countries.47
Figure 2 plots the real effective exchange rate, constructed as wholesale prices in Japan relative to wholesale prices in other countries converted into yen. We see there how the real rate rose around the time of the Korean War, consistent with the emphasis of Kosai (1989). It then fell by 2.8 per cent from the mid-1950s through the late 1960s, although the bilateral real rate against the U.S. dollar fell by about twice this amount.48 This is consistent with the hypothesis of emerging undervaluation, although the change in the real effective rate is small. Note also that while the real bilateral rate against the
43 As noted in Section 2 above.
44 Ackley and Ishi (1976, p.187) note that the 1963-64 and 1966-70 cyclical expansions were the first ones that were not marked by sharp and steady declines in net exports that ultimately resulted in their moving into negative territory. As Nakamura (1993, p.167) explains this, “the current account of the balance of payments stopped being a constraint on economic growth under the exchange rate of 360 yen per dollar because international competitiveness had improved enough” (authors’ translation).
45 See the appendix for details on the construction of these series.
46 Depending on the year in question. For purposes of the chain-weighted index, we updated the trade weights every five years.
47 Such as Greece and Thailand in 1953, Iran in 1957, the Philippines in 1962 and India in 1966.
14
United States continued to depreciate in the second half of the 1960s, the real effective rate did not. Figure 2 thus suggests that back-of-the-envelope calculations based on conveniently available U.S. and Japanese inflation rates, together with the constant bilateral exchange rate, tend to exaggerate these trends.
It can be argued that relative wholesale price indices adjusted for exchange rate changes understate the change in competitiveness, since wholesale prices are dominated by homogenous goods whose cost tends to be driven to equality by arbitrage. A better comparison for Japan would be the unit prices of exports relative to other industrial countries, since most Japanese exports were differentiated manufactures and the developed countries were Japan’s main competitors in markets for these products. Figure 3 confirms that export prices rose less rapidly than wholesale prices over the high growth period, especially in the early 1950s.
Figure 4 shows the unit value of exports since 1951 relative to the advanced economies.49 It indicates a substantial one-time improvement in Japan’s international competitiveness in the first half of the 1950s, as emphasized by Shinohara (1959). There is then a further improvement as the high-growth period proceeds. Japanese export prices fell by 7 per cent between 1960 and 1962, led by declines at twice this rate in metals, machinery and chemicals, three industries that had been the target of government-led
48 Again, the difference is explicable by the tendency for other countries to devalue against the dollar and therefore the yen, which lowered foreign prices when converted into yen.
49 The figures here are abstracted from “Comparative Statistics” compiled by the Bank of Japan, following the classification in the Monthly Bulletin of Statistics of the United Nations. The UN’s “Economic Class I” includes the U.S., Canada, Belgium, Luxembourg, France, West Germany, Italy, the Netherlands, Austria, Denmark, Norway, Portugal, Sweden, Switzerland, the U.K., Finland, Iceland, Ireland, Greece, Spain, Turkey, Yugoslavia, Japan, Australia, New Zealand and South Africa. The figures for each country reflect changes in export prices on each national currency basis. The data series for “developed countries” is aggregated after adjusting each country’s export price index for nominal exchange rates fluctuations against the dollar. 15
rationalization and modernization investment.50 After that, the unit value of Japanese exports held steady or rose slightly in periods when raw material prices were rising, but competitiveness continued to improve as a result of inflationary developments abroad and relatively rapid Japanese productivity growth and technical change. Overall, Japanese export prices fell by 29 per cent relative to those of its industrial-country competitors between 1951 and 1967. All this suggests growing undervaluation.
4. Exiting the Peg
Whether overvalued or undervalued, the government took the peg at 360 yen to the dollar as an “immutable condition.”51 Until Germany floated the deutschemark in May 1971 the issue of yen realignment received little attention; it was the subject of little systematic planning.52 Even then, Finance Minister Mizuta continued to insist that “the best policy choice is to achieve economic stabilization under the fixed exchange rate regime.”53 Those few economists and officials who did consider the possibility of revaluation tended to dismiss it as damaging to exports, investment and confidence.54
50 Krause and Sekiguchi (1976), p.420.
51 Kosai (1989), p.287 (authors’ translation). Thus, when German Economy Minister Ludwig Erhard visited Japan in 1958 and argued that the government should allow the currency to appreciate as part of a strategy of reducing government intervention in the economy, his remarks elicited a very strong negative reaction. The day following Erhard’s remark on the evaluation of the yen, Finance Minister Eisaku Sato commented at the press conference, “though Economy Minister Erhard said that one of the causes for extremely cheap prices of Japanese goods is the level of exchange rate, the current exchange rate of 360 yen per dollar is not undervalued. If Japan would revalue its exchange rate, Japanese trade industries could not cope with it.” (Nikkei Newspaper, 29 October 1958, authors’ translation). In the press, Erhard’s comments were interpreted as implying that a) Japan should abolish its capital controls, b) the country’s wage level is too low and should be raised, c) the exchange rate of 360 yen per dollar is undervalued and it should be revalued and d) Japan should expand domestic demand (see for example Hoshino 1958).
52 Bank of Japan (1986), p.303. Angel (1991, p.272) puts it more strongly, writing that “Expressions of doubt were banned absolutely within the MOF, and serious efforts were even made to prevent public discussion of the issue within the business community.”
53 Cited in Ministry of Finance (1992), p.374 (authors’ translation).
54 The prominent exceptions were the members of the Forum for Foreign Exchange Rate Policy, who recommended revaluation and shifting to a crawling peg in their report issued on 10 July 1971. However, 16
Fears that changing the exchange rate would have negative implications for employment growth and security in the manufacturing sector led some observers to warn that tampering with the exchange rate could even provoke political unrest.55
Of course, those advocating maintenance of the dollar peg had to offer alternatives for countering the chronic current account surplus, which caused tension with foreigners, and how to handle the inflationary pressures associated with capital inflows.56 The dominant recommendation was to relax exchange controls and trade barriers in order to encourage imports. For example, Takahashi (1971) stressed the need for import liberalization. The government in fact pursued this option by selectively removing foreign exchange controls and import quotas and relaxing regulations limiting outward FDI.57 Others like Shimomura (1971) plumped instead for fiscal stimulus to encourage imports and rebalance the current account. Still others worried that this option would only intensify inflationary pressure.58
Symptomatic of the deeply ingrained nature of inherited policy, Japan continued to peg the yen to the dollar, intervening in the foreign exchange market, for two weeks after President Nixon closed the gold window.59 Foreign reserves rose by nearly 50 per
the members of the forum recommended limiting the yen’s appreciation to 2 to 4 per cent a year, reflecting widespread worries about the negative repercussions of a large step revaluation.
55 Bank of Japan (1986), p.305. See also Takahashi (1971) and Hayami (1982).
56 The rise of trade tension with, inter alia, the United States, is of course another suggestive parallel between Japan in the 1970s and China today.
57 In June 1971. The relaxation of regulations included the liberalization of outward FDI in principle. The enhancement of liberalization of inward FDI was also in the list reflecting the criticism from abroad on the slow pace of liberalization (Ministry of Finance 1992, pp.368-370).
58 See for example Forum for Foreign Exchange Rate Policy (1971).
59 Note that this observation is consistent with the premise of this paper that Japan can be thought of as choosing to exit from the peg rather than being forced off it by events in the United States. Another not incompatible explanation for why the monetary authorities continued to place a ceiling on the dollar exchange rate for some weeks is wanted to allow banks to hedge their open positions. Japanese banks had accumulated dollar positions at the time of President Nixon’s announcement through the BOJ’s yen-denominated lending to banks to finance discounting foreign-currency denominated export bills. For the details of this scheme, see Bank of Japan (1986, pp.328-330). 17
cent between August 16th, the day following Nixon’s decision to close the gold window, and August 27th.60 After August 27th, the BOJ stopped purchasing dollars at the old price of 360 yen. The yen was allowed to appreciate, although the BOJ still intervened to slow its movement, accumulating more reserves. Once the yen reached 308 to the dollar, a 16.9 per cent bilateral appreciation, it was re-pegged.
[ 编辑 DadLU 在 08-04-23 04:36 ]
接上文，
拜托那个插队的mm删了把，我现在满腔怒火啊！ 吼！！！ :cn03: :cn03: :cn03:
Expectations were that revaluation would have deflationary effects on the economy, leading companies to postpone investment and run down their inventories. In addition, it was argued that abandoning the dollar parity that had been the anchor for policy since the late 1940s had negative effects on confidence.61
In response, the BOJ cut its discount rate by 50 basis points in December 1971 and by another 50 basis points in June 1972. A supplementary budget was passed for April 1971-March 1972, and a more expansionary stance was adopted for fiscal year 1972 with expenditure on general account up by 22 per cent and expenditure on public investment and lending through the FILP up by 32 per cent over the previous year.
The short-run impact of revaluation was relatively mild, reflecting the effects of all this monetary and fiscal stimulus. While exports fell by 0.1 per cent in the first quarter of 1972 (on a seasonally adjusted basis), they recovered strongly. In the fourth quarter of 1972, exports rose by 15.7 per cent from the same quarter of the previous year (in nominal yen-denominated terms). Exports of capital goods, such as general machinery, electrical machinery and transport equipment, rose by 28 per cent.62 The economy bottomed out in December 1971 according to the business cycle dates of the
60 This refers to the combined reserves of the Bank of Japan and the Ministry of Finance.
61 In the Bank of Japan’s retrospective analysis, the Japanese economy, which had begun to recover from the post-1969 recession in the summer of 1971, started to weaken again from the second half of August due to the effects of the so-called “Nixon shock.” Bank of Japan (1986), pp.357-358.
18
Economic Planning Agency. Real GNP grew by more than 10 per cent on an annualized basis in the first quarter of 1972, and the inflation rate bottomed on a CPI basis in September 1972, with a 4.7 per cent year-on-year increase.63 As early as March of that year there was discussion of whether the revaluation had been sufficient to prevent the current account from remaining in strong surplus, large amounts of foreign reserves from being accumulated, and inflationary pressures from intensifying.
With benefit of hindsight, Komiya and Suda (1983b) argue that Japan waited too long to exit from the peg and to move to a managed float. They write that “the government should have revalued the central rate of the Smithsonian Agreement or moved to floating at least by the autumn of 1972. [But] at that time the government and Bank of Japan considered the avoidance of revaluation as the primary policy goal and conducted macroeconomic policies accordingly. However, these policies accentuated imbalances among the major countries, especially between Japan and the United States. Furthermore, they generated the great inflation from 1973 to 1974…If the second shift to floating (in 1973) had been carried out half a year earlier and a tight monetary policy had been adopted, the outcome of policies could have shown a significant difference.”64
When the Smithsonian Agreement collapsed in early 1973, the yen was again allowed to float upward, this time to 265, at which point the BOJ again intervened to limit its fluctuation.65 Consistent with the advice tendered to China today, Japan’s float
62 The figure for transport equipment here does not include vehicles for personal use.
63 Subsequently it accelerated to 11.3 per cent in 1973 and 22.6 per cent in 1974, reflecting a combination of strong domestic demand, excess liquidity and the rise in oil and other commodity prices.
64 Komiya and Suda (1983b), pp.24, 31-2 (authors’ translation).
65 Komiya and Suda (1983b), p.41. It is interesting to note the parallels with Goldstein and Lardy’s (2003) proposal for China for a two-step exit from the yuan peg, the first step being a one-time revaluation against the dollar and the second step, after a period, being the transition to a managed float.
19
was heavily managed: the exchange rate was limited to a narrow range between 264 and 266 yen to the dollar through September 1973.
5. Capital Controls and the Forward Market
Throughout this period, exchange and capital controls remained in place. Restrictions on convertibility for purposes of transactions on current account were maintained into the 1960s.66 Each foreign currency transaction relating to imports required a license from MITI, while all foreign currency transactions for other purposes, such as transactions related to inward FDI, required one from the Ministry of Finance. Transactions on capital account were essentially suppressed. Up through 1963, the current account balance and changes in foreign reserves moved in lockstep, reflecting these pervasive controls on capital account.
The first step in liberalization was the establishment of yen accounts for nonresidents in 1960. The authorities then eliminated restrictions on transactions on current account in April 1964, accepting Article VIII of the Articles of Agreement of the International Monetary Fund. Inward investment was liberalized in 1968, but even then it was not entirely freed: significant restrictions on inward FDI remained into the 1970s.
The 1949 Foreign Exchange and Foreign Trade Control Law, which remained in effect until 1980, prohibited Japanese citizens from holding foreign exchange except with the permission of the authorities. There were also restrictions on the ability of Japanese financial institutions to maintain open positions in foreign exchange and on the ability of foreign financial institutions to take positions in domestically-issued yen-denominated
66 See Fukao (1990) for details. 20
securities.67 From 1968, restrictions on open positions were supplemented and then largely superceded by swap limits (yen conversion quotas under which the authorities set a ceiling for the net short spot positions of foreign currencies plus the amount of outstanding free yen liabilities to non-residents).68 This had the effect of limiting covered interest arbitrage by the banks, causing the forward rate to move more freely and making speculation more costly.69
Some of these restrictions were then tightened in the 1970s when the yen was unpegged, with the goal of limiting capital inflows. In 1972 controls were tightened on advances against exports, additional reserve requirements were imposed on increases in nonresident yen deposits, and limits were placed on nonresident purchases of Japanese securities.70 Then in 1973, when Japan’s current account swung into deficit first with the strong expansion of the economy and then the OPEC shock, controls on capital inflows were abolished while those on outflows were reinforced. When the yen strengthened in 1977, controls on the foreign currency deposits of residents were abolished, measures prohibiting residents from acquiring short-term foreign securities were eliminated, and a 50 per cent reserve requirement on increases in the yen deposits of nonresidents was instituted.71 Indicative of the extent of these capital controls, covered interest parity (the interest differential between the U.S. and Japan adjusted for the difference between spot and forward exchange rates) holds closely after December 1980, the moment of deregulation, but not before.72
67 The allowance of open positions for authorized banks was reduced twice in the mid-1960s to inhibit short-term capital outflows. Ministry of Finance (1992), pp.192-3.
68 Ministry of Finance (1992), p.212 and Fukao (1990), pp.119-121.
69 These ceilings were then transformed into limits on open positions in 1977 and abolished in 1984.
70 Fukao (1990), p.24.
71 Those reserve requirements were then doubled in 1978.
72 See Ito (1986).
21
For much of this period, authorization to participate in the foreign exchange market and the extent of permissible positions were subject to control by the Foreign Exchange and Foreign Trade Control Law and decrees. Only banks with approval from the government, so-called “authorized foreign exchange banks,” were allowed to book transactions on both the spot and forward markets. As mentioned before, the open positions of authorized banks were capped by decree.73 These regulations influenced forward market activities. So-called speculative transactions were prohibited, and only exporters and importers were said to have the opportunity of speculating and then only within the volume of their trade.74
The development of the forward foreign exchange market depended on the relaxation of foreign exchange controls and also on the incentive to invest in forward cover.75 The collapse of the Bretton Woods system therefore triggered the expansion of
73 In addition, residents were allowed to carry out foreign exchange transactions only when they were involved in trade and other limited transactions requiring foreign exchange. The law heavily regulated outward and inward portfolio investments. This restriction was called the “actual demand rule.”
74 Kawamura and Shibuichi (1961) explain this speculation as follows (p.11, authors’ translation):
“Apart from transactions based on actual demand (demand accompanied by actual trading such as exports and imports), speculative transactions can be carried out in the foreign exchange market. Forward transactions are sometimes used for purposes of speculation…. However, there is no room in Japan for speculation as described above because forward exchange contracts are allowed only on the basis of actual demand according to the Foreign Exchange Control Law. Transactions accompanied by actual demand can be used for speculation in certain cases. For example, exporters can expand their positions of forward foreign exchange well in advance when foreign currencies are expected to fall in the near future.”
Komiya and Suda (1983a) point out that exports and imports were mainly settled in U.S. dollars and exporters and importers had the chance to speculate by controlling the amount of their open positions (pp.147-152). Of course, there were ambiguities in the application of these measures that still left some scope for currency speculation. But precisely the same point can be made about China today, where exchange controls are comprehensive but still far from watertight.
75 Komiya and Suda (1983b) point out that trading volume of forward transactions increased as the spot market developed even under the fixed exchange rate system (p.23). It should be noted that even under the fixed exchange rate system, inter-bank and customer exchange rates were allowed to fluctuate within narrow bands after the deregulation of 1959, encouraging transactions in the forward market to a certain extent. Prior to September 1959, the so-called MOF exchange rate, by which authorized foreign exchange banks traded with the MOF, and customer exchange rates were fixed by decree of the Finance Minister. Deregulation allowed these rates to fluctuate between 0.5 per cent above parity, 360 yen per dollar, and 0.5 per cent below it (the band later expanded to 0.75 per cent in 1963). At the same time, regulation of the
22
foreign exchange transactions on both the spot and forward markets (Figure 5). This occurred only after the short-run disruption had passed, since the forward market stopped functioning with the yen’s floating in August 1971. Initially, the banks stopped quoting forward exchange rates for their customers, and trading volume on the inter-bank market collapsed. Exchange controls were then tightened in reaction to the collapse of the Bretton Woods System, resulting in a lack of opportunities for arbitrage and a drying up of the forward market. Forward market transactions resumed only in late October.76
In sum, Japan experienced gradual liberalization of its international financial transactions in the three decades ending around 1980. But even then there were extensive restrictions on the ability of households and financial institutions to engage in cross-border financial transactions and interest arbitrage. Exchange controls had been partially relaxed prior to floating of the yen, but they were still far from eliminated. And various controls were tightened or reimposed following the shift to managed floating.77
Some skeptics of the feasibility of a more flexible renminbi exchange rate argue that full currency convertibility, including on capital account, is a necessary prerequisite for greater exchange rate flexibility in practice. Others suggest that floating requires a deep and liquid forward market, and that China should not permit the renminbi to vary more freely and be determined by market conditions until there is significant additional progress in developing the inter-bank forward market in Shanghai.78 Japanese
forward exchange rate against the dollar was abandoned. (Forward rates against other currencies had been already deregulated.) Ministry of Finance (1992), pp.42-4.
76 Komiya and Suda (1983b), pp.22-23.
77 Revealingly, Fukao (1990) entitles a section of his paper “From 1973 to 1980: A Managed Float with Controls on Capital Movements.”
78 At the time of writing, several banks, including the Big Four banks, have started forward exchange transactions with their customers on a “pilot basis” (People’s Bank of China 2004b, p.56), but this is still far from a deep and liquid forward market. Among other things, the bank lack investment specialists with experience in managing foreign exchange exposures. Other foreign-exchange-market-related initiatives,
23
experience in the 1970s does not suggest that it is necessary to complete the task of liberalizing the capital account and developing the inter-bank forward market before moving to a significantly more flexible exchange rate. It suggests that banks and firms in China now, like those in Japan after 1973, should be able to tolerate a further increase in exchange rate volatility so long as the central bank continues to intervene to avoid very sudden movements in the currency’s level and sharp spikes in volatility.
6. Effects of Exiting the Peg
Efforts to identify the impact on Japan of exiting the peg are complicated by other disturbances hitting the economy around this time. By the early 1970s the high-growth period following World War II was coming to an end; after two decades of rapid export- and investment-led growth, much of the productivity gap vis-à-vis the United States had been closed. While the average annual rate of GDP growth decelerated from 12.1 per cent in 1960-69 to 7.5 per cent in 1970-73 and 3.8 per cent in 1973-85, it would be a mistake to attribute the entire shift to the change in exchange rate regime, since with the end of catch-up there would have been some slowdown in any case. In addition, the world economy grew very rapidly for two years following the exit, which may have cushioned the Japanese economy from any disruptive effects.
such as the enhancement of the Foreign Exchange Trade Centre, are largely limited to spot transactions. On 18 May 2005, inter-bank foreign currency trade was “formally introduced” in China. Previously, there was no trading platform for transactions between foreign currencies. The recent reform enables domestic financial institutions to trade eight pairs of foreign currencies in China; euro against U.S. dollar, Australian dollar against U.S. dollar, pound sterling against U.S. dollar, U.S. dollar against Swiss francs, U.S. dollar against Hong Kong dollar, U.S. dollar against Canadian dollar, U.S. dollar against Japanese yen and euro against Japanese yen. For the details of this reform, see People’s Bank of China (2005). While there also exists a market in nondeliverable forwards in Hong Kong, this is not accessible, by design, to domestic entities seeking to hedge foreign exposures.
24
As emphasized earlier, the motors of Japanese growth were exports and investment. One way of isolating the impact of the change in the exchange rate on growth is therefore to examine its effect on these two variables. Figure 6 juxtaposes capital investment in the manufacturing sector (as a share of GDP) with the real effective exchange rate. Investment shoots up once the real exchange rate comes down in the wake of the Korean War; it then declines in the early 1970s coincident with the appreciation of the real rate. The movement of manufacturing profits (current profits of the manufacturing sector as a share of GDP), upward after the early 1950s and then down after 1971, is consistent with the operation of this mechanism, with causality running from international competitiveness to manufacturing profitability and from there to capital formation in the manufacturing sector.
The problem is that other things were happening at the same time, some of which could have been affecting the real exchange rate, profitability and investment, with no causality running between them. The technological progress and industrial rationalization in the 1950s emphasized by Fujino (1988) and Kosai (1989) would have affected Japanese economic growth precisely by stimulating profitability and investment in manufacturing. The rapid growth of the world economy in 1971-3 could have supported investment and profitability in the face of a negative exchange-rate shock.
A more careful analysis of the connections between the exchange rate and investment is therefore required. A previous study along these lines is Miyagawa and Tokui (1994), who regress gross investment in manufacturing as a share of the capital stock on the real exchange rate and a vector of controls (a proxy for global demand, the ratio of wages to the cost of capital, and the ratio of intermediate input prices relative to
25
the cost of capital).79 They confirm that appreciation (depreciation) of the real effective exchange rates had a negative (positive) impact on Japanese investment in the 1980s.80 This is in line with the findings of other studies where real appreciation reduces export prices relative to costs, squeezing profitability and discouraging investment.
Here we extend their analysis back to the 1950s and consider not just the bilateral rate against the dollar but the real effective exchange rate. We estimate an equation of the form:
/kd /cw reer k /i1-t1-t 1-t1-t1-t1-ttδγβα+++= (1)
where
i: real capital investment in manufacturing
k: real capital stock in manufacturing
reer: real effective exchange rate
w: real wage in manufacturing
c: user cost of capital
d: world income
We lag the independent variables, assuming time to build. Consistent with a Cobb-Douglas production function, we enter the wage and cost of capital in ratio form. d, the GDP of major industrial countries, is our proxy for global demand for Japanese manufacturing products. It is normalized by the capital stock following the specification
79 Miyagawa and Tokui (1994) use effective exchange rates provided by J.P. Morgan in their estimates.
80 Other results are that global demand has a positive effect on investment, increases in the ratio of intermediate input prices to capital costs has a negative effect (as if capital and intermediate inputs are complements in production), and increases in the ratio of wages to capital costs has a positive effect (as if labor and capital are substitutes). Miyagawa and Tokui (1994) assume that trade industries are in a
26in Miyagawa and Tokui (1994). All variables are on annual basis and the sample period is 1955-1973.81
A problem with estimating this relationship in levels is that many macroeconomic time series exhibit unit roots. The augmented Dickey-Fuller test confirms that the real exchange rate and the factor price ratio have unit roots. In contrast, the hypothesis of a unit root can be rejected at the one per cent level for the investment ratio and world income normalized by the capital stock.82 Estimating this relationship in first differences consequently may introduce a problem of overdifferencing. We therefore estimated the equation both in levels and differences.
Table 4 shows the results. When the relationship is estimated in levels, the real effective exchange rate enters significantly, with a positive coefficient, indicating that appreciation had a negative impact on manufacturing investment. 83 The real wage relative to the user cost of capital is significant with a positive sign, indicating that a change in relative factor prices induces capital/labor substitution, other things equal.84 World income is also significant with a positive sign, confirming that an increase in world income had the expected positive effect on investment. The results using the differenced data, in the second panel, are consistent with these findings.
The effect of exchange rate on exports during high growth period is one of the major issues in the studies described in Section 3. Here, we estimate export function like that specified by Nakamura (1993):
condition of monopolistic competition where appreciation of their national currency leads to worsening the competitiveness and thus results in decline in share of national industries in the global market.
81 Each variable is transformed into index whose level of 1955 is 100. All data series are then expressed as logarithms.
82 The results of the DF-GLS test, which is supposed to have enhanced power in small samples, show the same results.
83 Note that real effective exchange rate here is defined as yen per unit of foreign currencies.
27
d reer extttγβα++= (2)
where
ex: real exports
reer: real effective exchange rate
d: world income
and α, β and γ are parameters to be estimated.
The results, again estimated over the period 1955-73 with annual data series, are in Table 5.85 Here estimates using the differenced data are more reliable given that we fail to reject the null hypothesis of a unit root in all series.86 The real effective exchange rate is significantly positive, confirming that the appreciation of the real exchange rate would have negatively affected Japanese exports. Global income is significant with positive sign, consistent with previous studies.
Simulating the impact of the exit on investment and exports requires an estimate of by how much the real exchange rate changed given post-1970 changes in the nominal exchange rate. Since regressing the real exchange rate on the nominal exchange rate would be tantamount to putting the nominal rate on both sides of the equation, we instead estimate a passthrough equation, regressing the WPI on the nominal effective exchange rate, foreign WPIs, and relevant controls. Since the real exchange rate is the ratio of domestic prices relative to foreign prices adjusted for the nominal exchange rate, we can use the results to estimate the impact of nominal appreciation on the real exchange rate (and hence on real variables like investment and export volumes).
84 As in Miyagawa and Tokui (1994).
85 All series are again transformed into logs prior to estimation.
86 Using both the augmented DF and DF-GLS tests.
28
The long-run relationship estimated here is of the form:
εtFortJPXwpi+Ω+
ββttewpi+=10 (3)
where wpiJP and wpiFor are the logs of the Japanese and weighted average of foreign wholesale price indices, e is the log of the nominal effective exchange rate (yen per unit of foreign currency) and X is a vector of controls (quarterly dummies, a linear time trend, real GDP, and the natural logarithm of the oil price, which figured importantly in Japanese price trends in the 1970s and 1980s). Note that this is only one of many forms in which this relationship is estimated; others have, for example, put the exchange rate on the left and domestic prices on the right and attempted to use it to track exchange rate fluctuations. Since in this period the exchange rate was tightly managed, while domestic wholesale prices were free to move, it makes more sense to treat the exchange rate as exogenous and domestic prices are endogenous.87
The equation is estimated using quarterly data for 1957-1990. The passthrough coefficient in Table 6 is in the range of 0.1-0.2. These estimates are relatively low: using data for a later period, Campa and Goldberg (2002) obtain a long-run passthrough coefficient for Japan of 0.8. But it is important to note that they consider import prices rather than wholesale prices, and wholesale prices are likely to move less, insofar as they include a large component of domestic prices which are less affected by exchange rate changes. Passthrough coefficients for other large economies such as the United States using relative wholesale prices also generate estimates on the order of 0.2.
87 And, of course, it is reasonable to assume that Japan was too small for its own inflation to significantly affect inflation in the rest-of-the-world aggregate.
29
A potential problem for the results in Table 6 is that we cannot reject the hypothesis of a unit root in the dependent and independent variables (see Table 7). However, these results will be consistent if the individual series are cointegrated. We therefore tested for the presence of unit roots in the residuals. The tests in Table 8 indicate that the null hypothesis of a unit root can be rejected.
The next step is estimating a short-run relationship that will allow for a more accurate characterization of the dynamics of adjustment. We use the residual from the preceding regression, denoted ut, as the error correction term in an equation designed to capture short-term dynamics.
ζμργγγttForttittiJPttiJPtXwpiewpiwpi+ΔΩ′++++=−−−=−−=−−=ΔΣΔΣΔΣΔˆ1121301113010131
(4)
We can solve this equation for the long-run effect of nominal appreciation on domestic prices (long-run passthrough), obtaining:
013111301−=−=Σ−Σ=titiγγφ
(5)
The estimates of this equation are in Table 9. The point is estimate of long-run passthrough is even smaller than before. However, this estimate is noisy due to the inclusion of some γ0 and γ1 coefficients that are not statistically distinguishable from zero. To correct for this we recalculated φ excluding γ coefficients that were not statistically significant at the 10 per cent level. The passthrough coefficient derived in this manner,
30
denoted φ’, is also reported in Table 9. These results are more stable across specifications. They suggest a long-run passthrough coefficient of 0.25.
A possible concern is that these results may be excessively influenced by the pre-1971 period when the yen was pegged to the dollar. Since the change in the nominal exchange rate was zero in this period as far as the bilateral rate against the dollar is concerned, passthrough from the exchange rate to prices will be zero as well, by construction.88 We took two approaches to addressing this problem. First, we restricted the sample to the period after the peg was abandoned. The coefficient on the nominal exchange rate turns out to be roughly the same as before, suggesting that the attenuation problem is not serious.89 Alternatively, we created a dummy variable for the period of floating and interacted it with the nominal exchange rate terms, adding this vector to the explanatory variables. This provides a direct test of the hypothesis that passthrough changed with the shift from pegging to floating. The long-run regression now becomes:
tFortPEGttJPtXDeewpiwpiεβθβ~*100+Ω+++=
(6)
where exchange rate passthrough is now given by φ” = β0 + θ0. The results confirm that passthrough was significantly higher in the floating period.90 But the additional effect is small; it has no noticeable impact on the overall level of long-run passthrough as estimated above.
88 Note that nominal effective exchange rates in the pre-1971 period fluctuated to some extent (for details, see Section 3 and Figure 1).
89 Arguments stemming from the Lucas Critique suggest that we might expect to find a higher passthrough coefficient in this later period insofar as floating made agents more aware of the possibility of exchange rate fluctuations. In practice this does not appear to be the case.
90 When a constant term is also included in the equation.
31
Recall that the nominal effective rate appreciated by 11.0 per cent between 1971Q2 and 1971Q4. Assuming a passthrough coefficient of 1/4, the induced appreciation of the real rate was three quarters of this, or 8.3 per cent. This is close to the actual appreciation of the real effective rate between 1971 and 1972 (see Figure 2).91 Between 1971Q1 and 1973Q1 the nominal effective rate rose by 26.2 per cent. Three quarters of this is 19.7 per cent. Again, this is rather close to the actual real appreciation observed in this period. We therefore take a real appreciation of 8.3 per cent as the short run impact and 19.7 per cent as the longer run impact of the exit.
We now are in a position to calculate the impact of the change in the nominal effective exchange rate on exports and investment. Taking the coefficients estimated on first-differenced data in Table 4 and 5, the short-run (1971-72) fall in exports is 8.6 per cent. The longer-run (1971-73) fall is 20.4 per cent. The impact on investment is larger: its level declines by 11.1 per cent between 1971-72 as a result of the 8.3 per cent real appreciation. The long-run effect is four times as large. These are substantial effects.
These results for exports and investment are consistent with one another. Figure 7 suggests that Japanese firms responded to the short-run change in the exchange rate by cutting their margins: they did not raise prices commensurately, which meant that they absorbed the much of the impact.92 That export prices do not rise as quickly as consumer
91 The actual appreciation on an annual average basis was 9.5 per cent.
92 In 1971-2, Japanese wholesale export prices declined by 3.2 per cent, while wholesale prices rose by 1.6 per cent, consistent with this distinction – and with the idea that Japanese exporters absorbed the exchange rate shock by cutting margins and thereby maintaining export market share. This response clearly varied by sector. The Economic White Paper of 1972 discusses the effects of yen’s revaluation on the competitiveness of exporters and price setting behavior of export sectors as follows (Economic Planning Agency 1972, pp.70-71, authors’ translation).
“Looking at changes in export prices (on a foreign currency denominated basis) and export volumes in the period from the shift to a floating exchange rate system (in August 1971) to March in this year (1972), export volumes of electric machinery and transport equipment increased even though
32
and wholesale prices in 1971-2 is consistent with this fact. In turn the decline in margins had a sharp negative impact on investment, which contributed to the recession that started in 1973.
Readers may wonder why a sharp slowdown in Japanese exports, investment and growth was not evident in 1972-3. Actual capital investment rose by almost 2 per cent between 1971 and 1973, while export volumes rose by 12 per cent. Our results suggest that the growth of exports and investment was mainly attributable to the rapid growth of world income, which rose by a cumulative 12 per cent in these two years. Normalized by the Japanese capital stock as we measure it, world demand increased by 34 per cent.93 Fortuitous conditions like these cannot be taken for granted. They caution that allowing a large and sudden appreciation following the exit from the peg could have a substantial negative impact on demand and economic growth.
7. Conclusion
Our analysis of Japanese currency experience in the 1970s suggests that a rapidly-growing, export-oriented economy can operate a floating exchange rate in the presence of capital controls and despite the absence of deep and liquid foreign currency forward
(foreign currency denominated) prices were raised enough to offset the impact of revaluation, reflecting strong competitiveness. The export volumes of steel and textiles fell due to the increase in export prices (on a foreign currency denominated basis) designed to partially absorb the effects of revaluation … Some industries with weak competitiveness in the global market such as chemicals and foods were forced to keep export prices (on a foreign currency denominated basis) to maintain export volume.
93 In the case of investment, other independent variables also moved in supportive directions. The ratio of wages to the cost of capital increased sharply, with both the numerator and denominator moving in supportive directions. Real wages increased by 22 per cent, while user cost of capital declined by 76 per cent due to the decline in the real interest rate. Real interest rates declined sharply: although the WPI increased by 15 per cent, the nominal bank lending rate rose only one per cent point. In this sense, monetary policy supported investment. In addition, public investment rose sharply in this period, suggesting that fiscal policy further supported investment.
33
markets so long as the central bank manages that float. Allowing the exchange rate to float enables the authorities to better tailor domestic money and credit conditions to local needs as the economy moves from a monetary and financial system that relies on administrative guidance to one in which interest rates guide the allocation of money and credit and in which the capital account of the balance of payments becomes increasingly open. But it is important for the country to manage its exchange rate with domestic conditions in mind and to avoid the kind of large real appreciation that would significantly compromise profits, damage investment, and jeopardize financial stability, especially in the absence of deep and liquid forward exchange markets on which foreign exposures could be hedged. If the floating rate is appropriately managed, however, it should not be necessary to wait on the further development of forward markets before allowing the currency to exhibit significantly greater flexibility.
For China this suggests starting with a modest band widening and limited increase in flexibility, similar to the policies in fact introduced in July 2005, and not acceding to foreign pressure for a large initial revaluation that might interrupt the growth of exports and depress investment. Other aspects of the comparison with Japan in the 1970s – the narrower profit margins of firms in the Chinese export sector, the more limited capacity of Chinese banks to cope with the financial consequences of a change in currency values, and the larger share of exports in Chinese GDP (which means that a smaller revaluation is needed to eliminate a current account surplus of any absolute size) – similarly point to the advisability of the Chinese authorities’ decision to start with a relatively small change in prevailing exchange rate.
34
Both that history and our assessment of current circumstances suggest that the kind of large real appreciation that might result from a premature shift to free floating should be avoided in the next step. This kind of large appreciation is advocated in some circles as helping to solve the problem of “global imbalances” – as China’s contribution to reducing the U.S. current account deficit. If Japan’s experience in the 1970s is any guide, a large real appreciation runs the risk of damaging profitability and investment and thus of slowing growth. A Chinese “contribution” to solving the problem of global imbalances that has this effect is in no one’s interest.
35
Table 1. Annual Growth Rates of National Income
Japan (GNP)
China (GDP)
1913-50
2.0
1913-40
4.0
+
1955-71
9.3
1950-78
4.9
1979-2003
9.4
1960s
10.1
1970s
4.7
1980s
4.0
9.7
1990s
10.4
2000-03
8.2
Notes: Japanese figure from 1913 to 1950 is calculated by splicing at 1931 the two series for GNP provided by LTES. GNP in 1945 is assumed to be two-thirds of 1944, following Maddison (2001). Post-war figures are estimated by the Economic Planning Agency. China’s data to 1978 are based on estimates by Maddison (2001), while those after 1979 are estimated by the IMF and the National Bureau of Statistics.
Sources:
Ohkawa, Kazushi, Nobukiyo Takamatsu and Yuzo Yamamoto (1974), Estimates of Long-Term Economic Statistics of Japan since 1868 volume 1, National Income, Toyo Keizai Shinposha.
Economic and Social Research Institute (ESRI), Cabinet Office, Statistics Information Site (www.esri.cao.go.jp).
IMF, International Financial Statistics.
National Bureau of Statistics, People’s Republic of China, China Monthly Economic Indicators.
36
Table 2. Exports and the Current Account
Japan
China
Export growth
Share of total US imports
Current account
Export growth
Share of total US imports
Current account
(annual average, %)
(%)
(billion US$)
(annual average, %)
(%)
(billion US$)
1950s
12.4
3.9
0.1
0.2
1960s
16.6
10.8
0.3
0.0
1970s
14.6
13.3
3.1
13.5
0.2
1980s
5.3
18.5
42.0
25.0
1.4
-1.8
1990s
2.3
15.6
99.4
23.4
6.4
12.5
2000s
3.5
10.7
114.0
22.4
10.4
24.4
Notes: Export growth is based on national currency-denominated figures. Share of US imports is calculated using US dollar-denominated figures. Share of US imports since 2000 is calculated using data for 2000-03. China’s current account figure for the 1980s is the average for 1982-89. That since 2000 is the average for 2000-02. Japan’s current account for the 2000s includes data through 2004.
[ 编辑 DadLU 在 08-04-23 04:43 ]
Sources:
IMF, Balance of Payments Statistics.
IMF, Direction of Trade Statistics.
IMF, International Financial Statistics.
Bank of Japan, Balance of Payments Monthly.
National Bureau of Statistics of China, China Statistical Yearbook 2001.
Yamazawa, Ippei, and Yuzo Yamamoto (1978), Estimates of Long-Term Economic Statistics of Japan since 1868, volume 14, Foreign Trade and Balance of Payments, Toyo Keizai Shinposha.
37
Table 3. Savings and Investment Rates
Japan
China
Savings rate
Investment rate
Savings rate
Investment rate
1950
29.9
16.2
1960
34.1
30.2
1970
40.5
35.1
1982
34.0
32.1
1990
38.0
35.2
2000
38.7
36.2
1st half of 1950s
27.5
20.1
2nd half of 1950s
29.0
24.4
1st half of 1960s
35.8
32.4
2nd half of 1960s
37.0
32.4
1st half of 1980s
34.5
33.2
2nd half of 1980s
35.9
37.5
1st half of 1990s
39.9
38.5
2nd half of 1990s
40.7
38.5
Notes: Japan’s rates are relative to GNP, China’s relative to GDP. China’s savings are the sum of gross capital formation and net current account. Figures for 1st half of the 1980s cover 1982-84.
Sources:
Ohkawa, Kazushi, Nobukiyo Takamatsu and Yuzo Yamamoto (1974), Estimates of Long-Term Economic Statistics of Japan since 1868 volume 1, National Income, Toyo Keizai Shinposha (Tables 1A, 6A).
National Bureau of Statistics of China, China Statistical Yearbook 2001 (Tables 3-11).
IMF, International Financial Statistics.
38
Table 4. Determinants of Investment
Annual Data
Estimated equation : /kd /cw reer k /i1-t1-t 1-t1-t1-t1-ttδγβα+++=
Figures in parentheses are t-statistics.
Levels
Constant term
Real effective exchange rate
Real wage / User cost of capital
World Income
Adjusted R2
DW statistic
-28.63
6.13
0.61
0.62
0.21
1.14
(-2.20)*
(2.46)*
(1.99)*
(2.35)*
Differences
Constant term
Real effective exchange rate
Real wage / User cost of capital
World Income
Adjusted R2
DW statistic
0.05
3.14
0.86
1.22
0.50
1.32
(0.60)
(1.88)*
(2.50)*
(2.04)*
Notes: 1.User cost of capital is calculated as:
wcP / d)-i ( P r +∗=π
r : user cost of capital
Pc : price index of capital goods (1952-1960), or that of investment goods (1960-1971)
Pw : wholesale price index (WPI)
i : banks’ lending rate (the averaged interest rate of banks’ new loans in December in each year)
π : percentage change in WPI
d : depreciation rate
2. ** denotes 1 per cent significance; * denotes 5 per cent significance.
3. For details on the construction of the real effective exchange rate, see data appendix.
4. Capital investment in the first quarter of 1955 is assumed as the same as averaged amount of second to fourth quarter. Capital stock at the end of 1952 is assumed as the same level as that of the end of March in 1953.
5. Real wage is total cash earnings per regular employee in manufacturing sector deflated by WPI.
6. World income is aggregated real GDP of eleven countries (see data appendix).
7. World income is normalized by the capital stock, following Miyagawa and Tokui (1994).
Source: see data appendix.
39
Table 5. Determinants of Exports
Annual Data
Estimated equation : d reer extttγβα++=
Figures in parentheses are t-statistics.
Levels
Constant term
Real effective exchange rate
World Income
Adjusted R2
DW statistic
-13.45
0.89
3.05
0.99
1.19
(-9.65)**
(3.12)**
(72.32)**
Differences
Constant term
Real effective exchange rate
World Income
Adjusted R2
DW statistic
0.04
1.04
2.27
0.35
2.08
(0.85)
(3.13)**
(2.23)*
Notes: 1. ** denotes 1 per cent significance; denotes 5 per cent significance.
2. Real exports are deflated by export price index on Long-Term Economic Statistics (LTES) basis (1953-1959) or by that of WPI basis (1960-1973).
3. For details regarding the construction of the real effective exchange rate, see data appendix.
4. World income is aggregated real GDP of eleven countries (see data appendix).
Source: See data appendix.
40
Table 6. Long Term Passthrough
Quarterly Data
(1)
(2)
(3)
(4)
0.142
0.142
0.136
0.137
Nominal Effective Exchange Rate (ln)
(0.030)**
(0.030)***
(0.028)***
(0.029)***
0.354
0.354
0.555
0.555
Weighted Foreign Price Level (ln)
(0.048)***
(0.048)***
(0.064)***
(0.064)***
0.114
0.114
0.061
0.061
Oil Price (ln)
(0.013)***
(0.014)***
(0.017)***
(0.017)***
0.067
0.067
0.160
0.161
GDP (ln)
(0.006)***
(0.006)***
(0.022)***
(0.022)***
Quarterly Dummies
No
Yes
No
Yes
Trend
No
No
Yes
Yes
Observations
136
136
136
136
R2
0.99
0.99
0.99
0.99
Standard errors in parenthesis.
significant at 10%; ** significant at 5%; *** significant at 1%.
41
Table 7. Unit Root Tests
Generalized Dickey-Fuller Test
All variables in logs
Variable
No Trend
Trend
Level
0.108
-15.40
WPI Japan
First Difference
-3.744**
-4.211***
Level
0.972
0.568
Nominal Effective Exchange Rate
First Difference
-5.332***
-5.209***
Level
0.578
-1.621
Weighted Foreign Price Level
First Difference
-2.616***
-2.980***
Level
0.312
-1.482
Oil Price
First Difference
-5.002***
-4.900***
Level
0.090
-0.522
GDP
First Difference
-1.019
-3.935***
All tests include a constant and four lags.
, , *** indicates unit root hypothesis can be rejected at 10%, 5%
and 1% significance levels respectively.
Table 8. Unit Root Tests
Generalized Dickey-Fuller Test
No Trend
Trend
Equation 1
-1.653
-3.086
Equation 2
-1.706
-3.115**
Equation 3
-1.692
-2.875
Equation 4
-1.741
-2.899*
All tests include a constant and one lag.
, , *** indicates unit root hypothesis can be
rejected at 10%, 5% and 1% significance levels
respectively.
42
Table 9. Passthrough Dynamics
(1)
(2)
(3)
(4)
Δpt-1
0.633
0.651
0.631
0.649
(0.092)
(0.093)***
(0.095)***
(0.096)***
Δpt-2
-0.101
-0.119
-0.122
-0.140
(0.108)
(0.111)
(0.112)
(0.114)
Δpt-3
0.025
0.036
0.013
0.023
(0.095)
(0.096)
(0.098)
(0.099)
Δet
0.091
0.089
0.087
0.085
(0.027)***
(0.027)***
(0.027)***
(0.028)***
Δet-1
-0.007
-0.004
-0.006
-0.003
(0.031)
(0.031)
(0.032)
(0.032)
Δet-2
-0.016
-0.022
-0.01
-0.016
(0.031)
(0.031)
(0.031)
(0.031)
Δet-3
-0.038
-0.033
-0.031
-0.026
(0.030)
(0.030)
(0.030)
(0.031)
Δp* t
0.283
0.267
0.275
0.259
(0.098)***
(0.099)***
(0.100)***
(0.101)***
Δp* t-1
-0.058
-0.032
-0.089
-0.063
(0.109)
(0.111)
(0.113)
(0.114)
Δp* t-2
0.219
0.200
0.203
0.184
(0.108)**
(0.109)*
(0.111)*
(0.112)
Δp* t-3
-0.101
-0.098
-0.125
-0.121
(0.092)
(0.093)
(0.094)
(0.096)
Δoilt
0.030
0.030
0.028
0.027
(0.009)***
(0.009)***
(0.009)***
(0.009)***
Δoilt-1
-0.021
-0.021
-0.014
-0.015
(0.009)**
(0.010)**
(0.009)
(0.009)
Δoilt-2
-0.015
-0.013
-0.011
-0.009
(0.011)
(0.011)
(0.011)
(0.011)
Δoilt-3
0.000
-0.001
0.005
0.004
(0.010)
(0.010)
(0.010)
(0.010)
Δgdpt
0.003
0.001
0.023
0.021
(0.059)
(0.059)
(0.064)
(0.065)
Δgdpt-1
0.090
0.094
0.107
0.11
(0.056)
(0.057)
(0.062)*
(0.062)*
Δgdpt-2
-0.000
-0.008
0.006
-0.002
(0.059)
(0.059)
(0.063)
(0.063)
Δgdpt-3
0.033
0.036
0.033
0.036
(0.054)
(0.054)
(0.058)
(0.059)
^
μt-1
-0.153
-0.154
-0.13
-0.130
(0.039)***
(0.039)***
(0.039)***
(0.039)***
φ
0.069
0.068
0.085
0.084
(0.106)
(0.109)
(0.100)
(0.103)
φ’
0.249
0.256
0.237
0.243
(0.102)**
(0.109)**
(0.101)**
(0.108)**
Quarterly Dummies
NO
YES
NO
YES
Trend
NO
NO
YES
YES
Observations
132
132
132
132
R2
0.75
0.75
0.74
0.74
Quarterly data, standard errors in parenthesis.
*significant at 10%; ** significant at 5%; *** significant at 1%.
43
Figure 1. Nominal Effective Exchange Rate (1949=100)
506070809010011012013014015049515355575961636567697173757779Nominal exchange rateagainst US dollarNominal effective exchange rateSources: See data appendix.(1949=100,yen/foreign currencies)
Figure 2. Real Effective Exchange Rate (1949=100)
506070809010011049515355575961636567697173757779Real exchange rate against US dollarSources: See data appendix.Real effective exchange rate(1949=100,yen/foreign currencies)
44
Figure 3. Prices (1949=100)
501001502002503004951535557596163656769717375Wholesale Price IndexExport Price IndexImport Price Index
Sources: Bank of Japan, Wholesale Price Index,
Ohkawa, Kazushi, Tsutomu Noda, Nobukiyo Takamatsu, Saburo Yamada, Minoru Kumazaki, Yuichi Shionoya and Ryoshin Minami (1967), Estimates of Long-Term Economic Statistics of Japan since 1868 volume 8: Prices, Toyo Keizai Shinposha.
Figure 4. Unit Price of Exports (1951=100)
5070901101301501701902102305155606567707275Source: Nakamura (1993), p.214, Table 61.Developed countriesJapan
45
Figure 5. Foreign Exchange Transaction Volume (1965-75)
02,0004,0006,0008,00010,00012,00014,00019651966196719681969197019711972197319741975spotforwardswap(million US$)
Note: Intraday trading volume, monthly averaged for December of each year.
Source: Nikkei Newspaper (Nihon Keizai Shinbun)
46
Figure 6. Real Effective Exchange Rate (REER) and Investment (1949=100; %)
50556065707580859095100495153555759616365676971737577790.01.02.03.04.05.06.07.08.09.010.0Capital investment of manufactuirng sector/GDP(right hand axis)REER(left hand axis)(%)(1949=100, yen/foreign currencies)Current profits of manufactuirng sector/GDP(right hand axis)
Sources: Ministry of Finance (1977), The Collected Data Series of Financial Statement Statistics of Corporations by Industry (Hojin Kigyo Tokei Kiho Syuran).
Economic and Social Research Institute (ESRI), Cabinet Office, Statistics Information Site (www.esri.cao.go.jp).
For REER, see data appendix. 47
Figure 7. Export Price Index/Input Price Index of Export Sectors
808590951001051101970197119721973(1970=100)
Sources: see data appendix.
48
49
Data Appendix
a) Nominal effective exchange rates (NEER)
Nominal effective exchange rates consist of yen exchange rates against the currencies of 17 countries: the United States, the United Kingdom, West Germany, Greece, Italy, the Netherlands, Sweden, Switzerland, Canada, Venezuela, Australia, India, Iran, South Korea, Thailand, the Philippines and South Africa. These are the countries that meet the following criteria; imports from Japan in 1960 exceed 10 billion yen, and both price data and foreign exchange rate data series are available from 1949 to 1990. (South Korea is an exception as its wholesale price index is not available in 1950. As figures in 1949 and 1951 are identical, we assume that the WPI in 1950 was unchanged from 1949.) The 17 bilateral rates are weighted by the value of imports from Japan, updated every five years. Note that the series depicted in Figure 1 excludes nominal exchange rate against South Korea as its extreme increase (its level in 1970 is 300 times larger than that of 1949) dominated the trends of other currencies.
b) Real effective exchange rates (REER)
Real effective exchange rates consist of yen exchange rates against the currencies of the same 17 countries. Each nominal exchange rate is normalized by the wholesale or producer price indices of Japan and the country in question, and weighted as above. For Japan’s wholesale price indices, domestic wholesale price index is used from 1960 while overall wholesale price index, which includes export and import prices, is used to 1959 due to data constraints. Note that the Figure 2 excludes the real exchange rate against South Korea for the same reason as above. Among the econometric results reported in the text, those which contain the data in the mid-1950s, namely investment and export equations, exclude South Korea from the weighted average to avoid the effects of the high-inflation episodes at the time.
c) World income (D)
We estimate world income by aggregating the GDPs of 11 countries: Australia, Belgium, Canada, France, Germany, Italy, Japan, Sweden, the Netherlands, the United States and United Kingdom with import value-based weights in 1955, 1960, 1965, 1970 and 1975. Imports of these countries account for about 60 per cent of world imports.
d) Sources for Table 4
Ministry of Finance, Financial Statements Statistics of Corporations (Hojin Kigyo Tokei), each issue; International Monetary Fund, International Financial Statistics, various issues; Direction of Trade Statistics, various issues; Maddison (2001); Mitchell (1998a, 1998b,1998c); Ministry of Trade and Industry, Annual Report of the Foreign Trade of Japan (Nihon Boueki Nenpyo), various issues; Bank of Japan (1986), One Hundred Year
50
History of the Bank of Japan, Volume of collection of historical materials; Bank of Japan (1987), Hundred-Year Statistics of Wholesale Prices in Japan.
e) Sources for Table 5
International Monetary Fund, International Financial Statistics, various issues; Direction of Trade Statistics, various issues; Maddison (2001); Mitchell (1998a, 1998b); Ministry of Trade and Industry, Annual Report of the Foreign Trade of Japan (Nihon Boueki Nenpyo), various issues; Bank of Japan (1987), Hundred-Year Statistics of Wholesale Prices in Japan; Yamazawa, Ippei and YuzoYamamoto (1978), Estimates of Long-Term Economic Statistics of Japan since 1868. Vol.14: Foreign Trade and Balance of Payments.
f) Sources for Figure 7
Input prices for export sectors are calculated using input price indices for industries whose export dependence in 1974-5 was more than 20 per cent: precision instruments, transport equipment, steel, general machinery, and textiles. Input price indices for these sectors are aggregated using their weights in the 1970 input price index. The source is: Bank of Japan, Price Indexes Annual, various issues.
51
References
Abegglen, James (1958), The Japanese Factory, Glencoe, Ill.: Free Press.
Ackley, Gardner and Hiromitsu Ishi (1976), “Fiscal, Monetary and Related Policies,” in Hugh Patrick and Henry Rosovsky (eds), Asia’s New Giant, Washington, D.C.: The Brookings Institution, pp.154-247.
Angel, Robert (1991), Explaining Economic Policy Failure: Japan in the 1969-1971 International Monetary Crisis, New York: Columbia University Press.
Association of Public and Corporate Bond Underwriters (Koshasai Hikiuke Kyokai) (1980), History of Public and Corporate Bond Market in Japan, Tokyo: Association of Public and Corporate Bond Underwriters (in Japanese, Nihon Koshasai Shijo Shi).
Bank of Japan (1976), The Financial System of Japan, Tokyo: Bank of Japan (in Japanese, Waga Kuni no Kinyu Seido)
Bank of Japan (1985), One Hundred Year History of the Bank of Japan, Volume 5, Tokyo: Bank of Japan (in Japanese, Nihon Ginko Hyakunenshi).
Bank of Japan (1986), One Hundred Year History of the Bank of Japan, Volume 6, Tokyo: Bank of Japan (in Japanese, Nihon Ginko Hyakunenshi).
Campa, Jose Manuel and Linda S. Goldberg (2002), “Exchange Rate Pass-Through into Import Prices: A Macro or Micro Phenomenon?” NBER Working Paper no. 8934 (May).
Dooley, Michael, David Folkerts-Landau and Peter Garber (2003), “An Essay on the Revived Bretton Woods System,” NBER Working Paper no.9971 (September).
Economic Planning Agency (1972), Economic White Paper of 1972, Tokyo: Printing Bureau of Ministry of Finance (in Japanese, Showa 47 Nen-ban Keizai Hakusho).
Eichengreen, Barry (2004), “Chinese Currency Controversies,” Asian Economic Papers (forthcoming).
Forum for Foreign Exchange Rate Policy (Kawase Seisaku Kenkyu-Kai) (1971), “A Recommendation for Adjustment of the Yen Rate by a Crawling Peg,” Shukan Toyo Keizai (special issue), 30 August (in Japanese, “En Reto no Kokizami Chosei ni tsuite no Teigen”).
Fujino, Shozaburo (1988), “The Balance of Payments of Postwar Japan, Part One: Overvaluation of Exchange Rate of 360 Yen,” Keizai Kenkyu 39, pp.97-108 (in Japanese, “Sengo Nihon no Kokusai Shushi”).
52
Fukao, Mitsuhiro (1990), “Liberalization of Japan’s Foreign Exchange Controls and Structural Changes in the Balance of Payments,” BOJ Monetary and Economic Studies 8, pp.1-65.
Goldstein, Morris and Nicholas Lardy (2003), “Two-Stage Currency Reform for China,” The Asian Wall Street Journal (12 September).
Hayami, Masaru (1982), Ten Years’ Experience with Floating: Voyage without a Chart, Tokyo: Toyo Keizai Shinposha (in Japanese, Hendo Soba Sei 10 Nen: Kaizu naki Kokai).
Horiuchi, Akiyoshi (1984), “Economic Growth and Financial Allocation in Postwar Japan,” Brookings Discussion Paper in Economics no. 18 (August).
Hoshino, Naoki (1958), “The Intention of Mr. Erhard,” Daiyamondo, 15 November, 1958 (in Japanese, Eaharuto no Shin’i).
Hutchison, Michael and Carl Walsh (1992), “Empirical Evidence on the Insulation Properties of Fixed and Flexible Exchange Rates: The Japanese Experience,” Journal of International Economics 32, pp.241-264.
Ishii, Kanji (1991), Japanese Economic History, Tokyo: University of Tokyo Press, Second Edition, (in Japanese, Nihon Keizaishi).
Ito, Takatoshi (1986), “Capital Controls and Covered Interest Parity Between the Yen and the Dollar,” Economic Studies Quarterly 37, pp.223-241.
Kawamura, Yozo and Tadashi Shibuichi (1961), The Practice of Forward Foreign Exchange, Tokyo: Gaikoku Kawase Boeki Kenkyu Kai (in Japanese, Sakimono Kawase no Jitsumu).
Kojima, Kiyoshi (eds) (1972), Structure and Development of Japanese Trade, Tokyo: Shiseido.
Komiya, Ryutaro (1988), Current Japanese Economy: Macroeconomic Perspectives and International Economic Relations, Tokyo: University of Tokyo Press (in Japanese, Gendai Nihon Keizai).
Komiya, Ryutaro and Miyako Suda (1983a), Contemporary International Finance: Theory, History and Policy; Theory Volume, Tokyo: Nihon Keizai Shinbun Sha (in Japanese, Gendai Kokusai Kinyuron, Riron-hen).
Komiya, Ryutaro and Miyako Suda (1983b), Contemporary International Finance: Theory, History and Policy, Volume for History and Policy, Tokyo: Nihon Keizai Shinbun Sha (in Japanese, Gendai Kokusai Kinyuron, Rekishi Seisaku-hen).
53
Kosai, Yutaka (1989), “Economic Policy During the Era of High Growth,” in Yasukichi Yasuba and Takenori Inoki (eds), Economic History of Japan, Volume 8: High Growth, Tokyo: Iwanami Shoten, pp.209-272, (in Japanese, “Kodo Seichoki no Keizai Seisaku,” Nihon Keizaishi).
Krause, Lawrence and Sueo Sekiguchi (1976), “Japan and the World Economy,” in Hugh Patrick and Henry Rosovsky (eds), Asia’s New Giant, Washington, D.C.: The Brookings Institution, pp.385-457.
Kure, Bunji and Kinzo Shima (1987), Deregulation of Interest Rates: New Edition, Tokyo: Yuhikaku (in Japanese, Kinri Jiyuka: Shinpan).
Maddison, Angus (2001), The World Economy: A Millennial Perspective, Paris: OECD.
Mason, Mark (1982), American Multinationals and Japan, Cambridge, Mass.: Harvard University Press.
Mikuni, Akio and R. Taggart Murphy (2002), Japan’s Policy Trap: Dollars, Deflation, and the Crisis of Japanese Finance, Washington, D.C.: The Brookings Institution.
Ministry of Finance (1976), The Financial History of Japan: The Allied Occupation Period, Volume 3, Tokyo: Ministry of Finance (in Japanese, Showa Zaisei Shi).
Ministry of Finance (1991a), History of Financial and Monetary Policies in Japan, 1952-73: Finance: Monetary Policy and Financial System, Part I, Volume 9, Tokyo: Ministry of Finance (in Japanese, Showa Zaisei Shi).
Ministry of Finance (1991b), History of Financial and Monetary Policies in Japan, 1952-73: Finance: Monetary Policy and Financial System, Part II, Volume 10, Tokyo: Ministry of Finance (in Japanese, Showa Zaisei Shi).
Ministry of Finance (1992), History of Financial and Monetary Policies in Japan, 1952-73: International Finance and External Economic Affairs, Part II, Volume 12, Tokyo: Ministry of Finance (in Japanese, Showa Zaisei Shi).
Ministry of Trade and Industry (1990), History of Trade and Industrial Policies, Volume 6, Tokyo: Tsusho Sangyo Chosa Kai (in Japanese, Tsusho Sangyo Seisaku Shi).
Mitchell, Brian (1998a), The International Historical Statistics: Americas 1750-1993, Basingstroke: Palgrave Macmillan.
Mitchell, Brian (1998b), The International Historical Statistics: Europe 1750-1993, Basingstroke: Palgrave Macmillan.
Mitchell, Brian (1998c), The International Historical Statistics: Africa, Asia & Oceania 1750-1993, Basingstroke: Palgrave Macmillan.
54
Mitsubishi Heavy Industry Ltd. (1967), History of Mitsubishi Shipbuilding Ltd., Tokyo: Mitsubishi Heavy Industry Ltd. (in Japanese, Mitsubishi Zosen Kabushikikaisha Shashi).
Miwa, Ryoichi (2003), “Postwar Democratization and Economic Reconstruction” in Nakamura Takafusa and Konosuke Odaka (eds), Economic History of Japan, Volume 3: A Dual Structure, Oxford: Oxford University Press.
Miyagawa, Tsutomu and Joji Tokui (1994), Economics of Strong Yen, Tokyo: Toyo Keizai Shinposha (in Japanese, Endaka no Keizaigaku).
Nakamura, Takafusa (1993), The Japanese Economy: Its Growth and Structure, Tokyo: University of Tokyo Press, Third edition (in Japanese, Nihon Keizai Shi: Sono Seicho to Kozo).
Nakamura, Takafusa (2003), “The Age of Turbulence: 1937-54,” in Nakamura Takafusa and Konosuke Odaka (eds), Economic History of Japan, Volume 3: A Dual Structure, Oxford: Oxford University Press.
Odaka, Konosuke (1989), “The Trace of Growth (2),” in Yasukichi Yasuba and Takenori Inoki (eds), Economic History of Japan, Volume 8: High Growth, Tokyo: Iwanami Shoten (in Japanese, “Seicho no Kiseki(2),” Nihon Keizaishi).
Odaka, Konosuke (1999), “‘Japanese-Style’ Labour Relations,” in Tetsuji Okazaki and Masahiro Okuno-Fujiwara (eds), The Japanese Economic System and Its Historical Origins, Oxford: Oxford University Press.
Ohkawa, Kazushi and Henry Rosovsky (1973), Japanese Economic Growth, Stanford: Stanford University Press.
Okazaki, Tetsuji (1993), Japanese Industrialization and the Steel Industry: A Comparative Institutional Analysis of Economic Development, Tokyo: University of Tokyo Press (in Japanese, Nihon no Kogyoka to Tekko Sangyo: Keizai Hatten no Hikaku Seido Bunseki).
Okazaki, Tetsuji (1999), “Corporate Governance,” in Tetsuji Okazaki and Masahiro Okuno-Fujiwara (eds), The Japanese Economic System and Its Historical Origins, Oxford: Oxford University Press.
Organization for Economic Growth and Cooperation (1994), Employment Outlook, Paris: OECD.
People’s Bank of China (2004a), China Monetary Policy Report: Quarter One, 2004, Beijing: China Financial Publishing House.
55
People’s Bank of China (2004b), China Monetary Policy Report: Quarter Four, 2004, Beijing: China Financial Publishing House.
People’s Bank of China (2005), “Foreign Currency Trading Formally Introduced to the Inter-bank Foreign Exchange Market,” Press Release, http://www.pbc.gov.cn/english.
Shimada, Haruo, Toyoaki Hosokawa and Atsushi Seike (1982), “The Analysis of the Adjustments of Wage and Employment,” Keizai Kenkyu, no.84 (in Japanese, Chingin oyobi Koyo Chosei Katei no Bunseki).
Shimomura, Osamu (1971), “Keeping the Parity Should Be the Principle of Economic Policy,” Shukan Toyo Keizai (special issue), 30 August (in Japanese, Heika Iji koso Keizai Seisaku no Kihon dearu).
Shinohara, Miyohei (1959), “Liberalization and Exchange Rate of 360 Yen,” Ekonomisuto, 10 November (in Japanese, “Jiyuka to 360 Yen Reto”), reprinted in Japan and the World Economy (1989), pp.22-30, Tokyo: Chikuma Shobo (in Japanese, Sekai Keizai to Nihon).
Suzuki, Yoshio (eds) (1987), The Japanese Financial System, Oxford: Clarendon Press.
Tachi, Ryuichiro and Ryutaro Komiya (1960), Under-Liquidity and Monetary Policy in Japan,” Economic Studies [Keizai Kenkyu], vol.11, no.3, pp.288-295.
Takahashi, Kamekichi (1971), “Yen’s Revaluation,” Asahi Newspaper, 4 May 1971 (in Japanese, En Kiriage).
Takamura, Naosuke (1971), The Introduction for the History of Japanese Cotton Spinning Industry, Tokyo: Hanawa Shobo, (in Japanese, Nihon Bosekigyo-shi Josetsu).
Toyota Motor Corporation (1987), Unlimited Creativity: Fifty Year History of Toyota Motor Corporation, Toyota: Toyota Motor Corporation (in Japanese, Sozo Kagiri Naku: Toyota Jidosha Goju-nen Shi).
Trezise, Philip and Yukio Suzuki (1976), “Politics, Government, and Economic Growth,” in Hugh Patrick and Henry Rosovsky (eds), Asia’s New Giant, Washington, D.C.: The Brookings Institution, pp.753-811.
Yasuba, Yasukichi and Takenori Inoki (1989), “An Overview: 1955-80,” in Yasukichi Yasuba and Takenori Inoki (eds), Economic History of Japan, Volume 8: High Growth, Tokyo: Iwanami Shoten, pp.1-56 (in Japanese, “Gaisetsu, 1955-80 Nen,” Nihon Keizaishi)
[ 编辑 DadLU 在 08-04-23 04:42 ]
[ 编辑 DadLU 在 08-04-23 04:45 ]
英文部分 找的好苦 
第一页，第一篇没完，这里弄完
No. C2005002 2005-01
人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
No. C2005002 2005 年1 月11 日
2
人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
No. C2005002 2005 年1 月11 日
摘要：本文运用行为均衡汇率模型对人民币均衡实质汇率和汇率失调程度进行了实证研究，样本区间为1991 年1 季度–2004 年3 季度。本文的主要发现是：（1）从1994 年4 季度起人民币均衡实质汇率处于不断升值的状态，其背后的主要驱动力量是我国制造业劳动生产率的快速上升和经常项目盈余导致的净对外资产余额的不断增加；（2）1990 年代以来，人民币实际实质汇率在大部分时期偏离均衡实质汇率轨迹，表现为人民币汇率的失调。其中，1992 年2 季度–1994 年4 季度为人民币汇率低估时期， 1995 年1季度—1999 年2 季度为人民币汇率高估时期，而1999 年3 季度往后的时期人民币汇率重新转为明显的低估，并且低估程度有进一步扩大的趋势。本文对人民币汇率失调的原因分析表明1997 年以来事实上的钉住美元的汇率政策是造成人民币汇率失调的一个主要的宏观政策因素。因此，本研究的政策含义是，从应对人民币汇率失调的角度，一个更为灵活的人民币汇率制度将更有利于中国经济的健康发展。
关键词：人民币，均衡实质汇率，汇率失调，行为均衡汇率模型
3人民币均衡汇率与汇率失调：1991-2004
施建淮
北京大学中国经济研究中心
余海丰
红塔证券股份有限公司资产管理总部
摘要：本文运用行为均衡汇率模型对人民币均衡实质汇率和汇率失调程度进行了实证研究，样本区间为1991 年1 季度–2004 年3 季度。本文的主要发现是：（1）从1994 年4 季度起人民币均衡实质汇率处于不断升值的状态，其背后的主要驱动力量是我国制造业劳动生产率的快速上升和经常项目盈余导致的
净对外资产余额的不断增加；（2）1990 年代以来，人民币实际实质汇率在大部分时期偏离均衡实质汇率轨迹，表现为人民币汇率的失调。其中，1992 年2 季度–1994 年4 季度为人民币汇率低估时期， 1995 年1季度—1999 年2 季度为人民币汇率高估时期，而1999 年3 季度往后的时期人民币汇率重新转为明显的低估，并且低估程度有进一步扩大的趋势。本文对人民币汇率失调的原因分析表明1997 年以来事实上的钉住美元的汇率政策是造成人民币汇率失调的一个主要的宏观政策因素。因此，本研究的政策含义是，从应对人民币汇率失调的角度，一个更为灵活的人民币汇率制度将更有利于中国经济的健康发展。
关键词：人民币，均衡实质汇率，汇率失调，行为均衡汇率模型
一、引言
近年来，人民币汇率问题成为学术界、商界、和政策当局热烈讨论的话题。首先是东亚货币危机期间，在亚洲主要货币大幅贬值的背景下，国际金融市场和国内投资者都预期人民币将会贬值，经济学家们也就人民币是否应该贬值展开了激烈的争论。而近两年来，在中国经济快速增长，经常项目顺差不断扩大和外汇储备激增的背景下，国际上要求人民币升值的呼声不断，国内学术界和政策当局关于应否调整人民币汇率和改革人民币汇率制度的讨论也趋于激化。
理论研究表明，汇率失调（exchange rate misalignment），不论是汇率的高估还是低估，都会使经济付出福利和效率方面的代价。而大量的实证研究也表明，汇率失调会对经济产生持久、深远的影响。如Edwards & Savastano（1999）指出“持续的高估应被视为是货币危机的一个非常重要的先兆；汇率的持续失调通常是与经济在中长期内的低迷联系在一起的”，Cottani et al.（1990）运用24 个发展中国家的面板数据发现，人均收入、出口、净投资和农业的增长都与汇率失调有明显的负相关关系。事实上，从事后的角度，人们一般认为，在东亚货币危机期间人民币出现了一定程度的高估。这种高估对当时的中国经济产生了显著的影响，如出口，1998 年几乎没有增长，1999 的增长率也只有6.1%；对国外直接投资的影响更为明显，1998 年实际利用外资没有增长，1999 年出现了11%的负增长，2000 年的增长可以忽略不计。
由于汇率失调会对经济产生深刻影响，人民币汇率是否失调？失调程度如何？便是中国政策当局宏观经济管理中需要做出判断的重要问题，也是应否调整人民币汇率和改革人民币
4
汇率制度问题的核心。为了回答这些问题，我们需要知道人民币的均衡汇率，以作为判断人民币汇率是否失调，失调程度如何的标准。需要指出的是，在本文中我们关心的“人民币汇率”是人民币实质汇率（经由名义汇率转换的外国一般价格水平与本国一般价格水平之比），因为影响贸易和投资的是实质汇率，它是开放经济中最重要的相对价格之一。然而由于人民币的均衡实质汇率是不可观察的，需要我们使用科学的方法来测算。本文的目的是就人民币的均衡实质汇率和汇率失调程度进行测算，以回答人民币汇率是否失调，失调程度如何的问题。具体而言，我们关心的问题有两个，一是人民币均衡实质汇率行为与经济基本面变量之间有怎样的关系，二是人民币汇率失调情况究竟如何（东亚货币危机期间人民币是否高估，高估程度如何；当前人民币是否被低估，低估程度又如何）。并试图分析人民币汇率失调背后的经济原因。传统上经济学家们常常运用购买力平价（PPP）原理测算均衡实质汇率以及汇率失调程
度。购买力平价原理主张名义汇率是由国内外价格水平的相对变动来决定的，因而隐含着均衡实质汇率为常数的结论。该方法首先确定一个基期1，然后将该时期的实质汇率作为所有考察时期内实质汇率均衡值的估计，任一时期实际实质汇率与基期实质汇率的偏离便作为汇率失调程度的测度。这种基于购买力平价的测算方法虽然简单和直接，但并非评估汇率失调程度的好方法。一个原因是，购买力平价在现实中并不成立，如果说购买力平价的基础是商品的国际套利，由于一般价格水平计算中包括了不可贸易品，那么就没有理由认为购买力平价在现实中会成立从而构成均衡实质汇率测算的基础2。此外，这种方法无法说明均衡实质汇率本身的变化。现实中经济的基本面总是会发生变化的，政策当局感兴趣的焦点是基本面的变化是如何改变实际的和均衡的实质汇率，从而改变汇率失调的程度的。自Williamson（1985）提出基本均衡汇率 fundamental equilibrium exchange rate, 简称FEER）概念以来，许多旨在测算均衡实质汇率和汇率失调的正式方法被相继开发出来，这些方法的一个主要优越之处是能够说明均衡实质汇率本身的变动。在这些正式方法中，以Williamson（1994）为突出代表的一类方法（FEER 方法）将重点放在宏观平衡上，通过要求实质汇率与宏观平衡相一致来确定实质汇率应该达到的水平，并将该值定义为均衡实质汇率（称为基本均衡汇率）。这里，宏观平衡指经济处于充分就业和低通货膨胀（内部平衡）以及经常项目反映了可持续的净资本流动（外部平衡）这样一种理想状况。均衡汇率概念之所以被冠以“基本”一词是因为它抽象掉了短期经济因素，仅仅关注于中长期经济状况3。FEER 方法的主要缺陷是可操作性差，并由于涉及到大量与经常项目和资本项目有关的参数设定（将参数校准在充分就业和可持续的净资本流动水平），使得该法得到的估计结果对模型参数的设定比较敏感。此外，由于FEER 是规范意义上的均衡汇
率概念，FEER 方法代表了均衡汇率的规范经济学分析方法，一些被证明对实际实质汇率行为有影响的经济变量并没有包括在FEER 的计算框架中，因而FEER 方法计算的均衡实质汇率在实证意义上是否存在，即计算的均衡实质汇率是否反映了那些实质汇率的决定因素在中期的影响，是不明确的。与FEER 方法相对照，另一类测算均衡实质汇率和汇率失调的方法将重点放在实际实质汇率行为本身，通过对实质汇率有影响的相关经济变量来解释实际观察到的名义汇率和实质
1 该时期的经济被判断为处于宏观平衡状态。
2 已有的实证文献已充分证明：实质汇率的时间序列不是平稳的，并且当使用非常长的样本或使用面板数据而发现它是均值反转时，其调向用相对价格显示的均衡路径的速度也是非常慢的。因此难以用ppp 概念来解释实际实质汇率对其均衡水平的持久偏离。
3 Wren-Lewis（1992）将FEER 方法定义为“一种计算与中期宏观平衡相一致的实质汇率的计算法”。对于FEER 方法而言重要的是中期宏观平衡的概念，该方法实际上并没有对如何进行均衡实质汇率的计算规定统一的模型，实质汇率计算既可以通过大规模联立宏观经济计量模型进行，也可以通过一个局部均衡模型来计算。汇率的运动。这类方法故而被称为行为均衡汇率法（behavioral equilibrium exchange rate,简称BEER，参见Clark & MacDonald，1999）。BEER 法运用近年来计量经济学发展起来的协整（co-integration）技术，从统计学意义上发现实质汇率和早先文献识别出的各种中长期汇率决定因素之间的协整关系，以此作为确定均衡实质汇率和评估汇率是否失调的基础4。由于BEER 法只涉及到单一方程约化型（reduced form）模型的估计，较之FEER 方法具有可操作性强的优点。因此，近年来BEER 法被广泛应用于均衡实质汇率测算和汇率失调问题的实证研究（例如参见Baffes et al.，1999，Clostermann & Schnatz，2000，Maeso–Fernandezet al.，2002）。对于研究发展中国家的汇率失调问题而言，BEER 法具有进一步的优势：尽管在发展中国家的实证分析中存在诸如样本量小，数据质量不高以及经济结构不稳定等问题，但是BEER 法分析的结果常常能够如理论预言的那样发现实质汇率与基本面变量之间的协整关系，并且估计出的协整方程常常能够重现通过其它方法识别出的汇率失调情况
（Montiel，1999a）。关于人民币均衡实质汇率的测算，自东亚货币危机以来，也出现了一些规范的文献，如张晓朴（1999）、Zhang（2001）、林伯强（2002）和张斌（2003）等。Zhang（2001）和林伯强（2002）使用的是年度数据，为了增加样本的长度，他们都采用了1990 年以前的数据，甚至把样本的起点放在上个世纪50 年代，从而样本包含了各种各样的汇率体制。事实上，在1980 年代以前，价格在本质上只是政府计划的一个工具，并没有成为反映市场供需的信号。所以，对利用这些数据计算出来的结果，其可信性值得商榷。其次，张晓朴（1999）、Zhang（2001）和林伯强（2002）所使用的数据都是2000 年以前的，这样，他们的研究结果只能提供东亚货币危机期间人民币是否高估的信息，而对目前人民币是否低估不能给出回答。最后，他们对均衡实质汇率和汇率失调的情况以及背后的经济原因没有进行充分的分析，例如Zhang（2001）实际上只讨论了人民币汇率当前失调的情况5。与Zhang（2001）和林伯强（2002）不同，张晓朴（1999）和张斌（2003）使用季度数据进行人民币均衡实质汇率测算，这在数据质量上是一大改善。例如张斌（2003）采用1992 年1 季度—2002 年4 季度的样本数据，应用Baffes et al.（1999）的模型和方法估计了人民币的均衡实质汇率和汇率失调的程度。不过，他选择的基本面变量均为流量，所以其估计仅仅反映了流量均衡而忽略了资产存量的影响。此外，张晓朴（1999）和张斌（2003）还都将一些名义变量作为决定实质汇率的基本面变量（前者选择了货币供给量，后者选择了世界出口品价格），这种做法也值得商榷，因为正如超调模型（Dornbusch，1976)指出的，在长期，名义冲击不影响实质汇率6。而且货币供给只是一种政策工具，将其作为基本面变量是不合适的。本文的目的是：采用1991 年1 季度—2004 年3 季度更新的季度数据，运用行为均衡汇率（BEER）模型估计人民币的均衡实质汇率和汇率失调程度，并深入分析人民币汇率失调背后的经济原因。本文剩余部分的安排如 ：第二部分阐述我们估计人民币均衡实质汇率的计量模型和变量选择；第三部分是模型的检验和估计；第四部分对计量结果进行分析说明；第五部分给出总结性评论。
4 如果实质汇率与经济基本面变量是协整的，那么实质汇率在长期内就具有均值反转的性质，而协整方程的均值就可以视作是长期实质汇率的均衡值。
5 参见下一节关于当前失调的定义。
6 Obstfeld and Rogoff (1995)试图通过价格粘性证明实质汇率滞后现（real exchange ratehysteresis）的存在，从而证明货币冲击在长期对实质汇率有影响，但实证研究没有支持其结论（Rapach，2001）。
6
二、计量模型和变量选择
BEER 方法通过估计一个解释实际实质汇率行为的约化型方程来确定均衡实质汇率水平
和汇率失调程度，这种约化型方程的线形形式可表述如下：
t t t t q Z T â è å Œ Œ = + + （1）
其中t q 表示实际观察的实质汇率（用对数形式表示）， t Z 是那些对中长期实质汇率有影响的经济基本面变量（对数形式或比率）的当前值向量， t T 是那些对短期实质汇率有影响的短期和一时性变量（如泡沫，政策当局的发言等）构成的向量， , â è是约化型（reduced form）参数向量， tå是随机扰动项。这里我们可以看出BEER 方法与FEER 方法的另一个区别：FEER是一个中长期概念，而BEER 则更为宽泛，原则上BEER 方法也可用于解释实质汇率的周期性
运动。例如原则上可以将t t Z T â è Œ Œ + 作为短期均衡实质汇率的估计。不过由于这种短期均衡
[ 编辑 DadLU 在 08-04-23 05:03 ]
swzlhzyy于2008-04-23写道:
楼主 我是很想看的，很佩服你的精神，但不能给个链接什么的阿
实在看不下去，这格式太乱了，文章都变形了
不然你把附件发出来吧
请不要插毫不？哭
实质汇率的估计中包括了一些泡沫因素和许多一时性因素的影响，准确估计短期均衡实质汇
率既不太现实也没有多大的政策意义。
Clark & MacDonald（1999）定义：
t t q Zâ Œ Œ = （2）
为当前均衡实质汇率（current equilibrium rate）,即该均衡实质汇率是利用经济基本面
变量的当前值（current values）计算出来的，并将实际的实质汇率t q 与当前均衡实质汇
率t qŒ 之间的差定义为当前失调（current misalignment）。当前失调可表述为：
t t cm q qŒ = - （3）
由于经济基本面变量的当前值本身也可能偏离其长期均衡水平，因此Clark &
MacDonald（1999）进一步定义
* *
t t q Z⌠= （4）
为长期（或持久）均衡实质汇率（permanent equilibrium rate），其中*
t Z 为经济基本面变
量的长期均衡值向量。将实际的实质汇率t q 与长期均衡实质汇率*
t q 之间的差定义为长期（或
持久）失调（permanent misalignment）7，长期失调可表述为：
*
t t pm q q = - （5）
BEER 方法在很大程度上强调的是实证意义，其经济基本面变量集的选择多少有些特定
7 Clark & MacDonald（1999）称之为总失调（total misalignment）。
7
或随意性（ ad hoc），因此BEER 方法的运用依赖于选择适当经济基本面变量集的理论指导。自从Edwards（1988）以来，已经出现了大量关于基本面变量选择的理论（见Montiel，1999a的一个综述），它们构成实际运用BEER 方法的基础。例如，Clark & MacDonald（1999）运用BEER 方法估计了德国马克、日元和美元实质有效汇率方程，基于Faruqee（1995）和MacDonald（1997）的研究，Clark & MacDonald（1999）认为就其研究目的而言，实质利差、贸易条件、非贸易品与贸易品的相对价格比、净对外资产以及本国与外国的政府债务比足以构成影响实质汇率的基本面变量集。有关基本面变量选择的理论成果与检验经济变量之间是否存在均衡关系的协整技术两者的结合保证了BEER 方法的有效性。在选取决定人民币均衡实质汇率的经济基本面变量时，我们考虑以下三个因素：一是理论模型所建议的变量，主要根据BEER 方法的已有文献所给出的建议（如Faruqee（1995）、Clark & MacDonald（1998）和Montiel（1999b）等）；二是数据的可得性；三是我国的具体国情。综合考虑以上三点，本文选取的经济基本面变量有：贸易条件（TOT）、非贸易品与
贸易品的相对价格比（TNT）、净对外资产（NFA）和反映贸易政策的变量（TRADE）8。下面
我们对各变量的意义及数据来源作一简单说明。
人民币实质汇率：人民币实质汇率用实质有效汇率（real effective exchange rate, 简
称REER）来衡量。用自然对数的形式表示即为：
7
1
ln( ) ln( / ) i i i
i
reer REER w S CPI CPI
=
= =‡”
其中CPI , i CPI 分别指我国和各贸易伙伴的消费者价格指数，我们考虑的贸易伙伴依次为
美国、日本、德国、英国、韩国、香港和台湾地区，它们与我国的贸易额占我国总进出口的
比重平均高达65%左右，在1997 年之前更是达到了70%。i S 表示我国与i 国（或地区）的名
义双边汇率。注意，我们这里采用间接标价法，即人民币的外币价格，名义、实质有效汇率
的上升（下降）意味着人民币的升值（贬值）。由于在1994 年以前我国实行的是双轨的汇率
制度（官方牌价与调剂市场价格共存），但外汇调剂市场承担了80－85%左右的交易量，所
以本文对1991 年1 季度－1993 年4 季度期间的名义汇率数据使用外汇交易市场的人民币价
格。i w i ＝ 贸易伙伴与我国的贸易额/7 个贸易伙伴与我国的贸易额总额。
贸易条件（TOT）：贸易条件被定义为出口价格与进口价格之比，它被用来描述一个国家
的贸易品在国际市场上的竞争力9。这里我们采用相对有效贸易条件指标，其定义为我国与
“世界”的贸易条件之比，后者通过主要贸易伙伴的贸易条件几何加权平均得到，用对数的
形式表示为：
7
1
ln ln( / ) ln( / ) i i i
i
tot TOT EX IM w EX IM
=
= = -‡”
8 这些变量都是研究均衡实质汇率问题的经济学家们一致认同的基本面变量（参见Williamson 为其1994 年
编辑著作所写的引言）。没有考虑国内外的实质利率差的原因是：到目前为止我国对资本的国际流动实行了严格管制，实质利率的影响不太显著，此外，经济学家们对是否将利率作为基本面变量尚有争议。
9 绝大多数的文献都不加说明的认为，一国贸易条件的改善（恶化）会导致该国货币的升值（贬值），但需要指出，贸易条件的改善有两个效果：一是“收入效应”，出口品价格的相对上升意味着实际收入的增加，从而更多地需求非贸易品；一是“替代效应”，进口品价格的相对下降，会增加对进口品（包括中间品和最终产品）的需求。前者推动了非贸易品价格的上升，从而有助于国内价格的上涨，而后者则有利于国内价格的下降。因此，贸易条件的改善对国内价格水平的影响是不定的，从而对实质汇率的影响也是不定的。
8
其中, EX IM分别指出口、进口价格指数。
非贸易品与贸易品的相对价格比（TNT）：该指标是一个衡量本国与外国生产率增长差
异的较为间接的指标，实证分析中也有用实质GDP 与全部劳动人口的相对比率这一更直接的指标来衡量本国与外国生产率增长差异的。直接指标试图抓住生产率增长的趋势，而间接指标则试图抓住贸易品与非贸易品部门生产率增长的差异从而体现Balassa-Samuelson 效应10。间接指标被广泛地运用于均衡实质汇率的实证研究中（如Chinn（1999），Clark &
MacDonald（1999），Clostermann & Schnatz（2000））。理论上，该指标需要使用非贸易品与贸易品的价格指数去计算。但是，在实际操作时，我们无法得到这两个价格指数。遵从文献的一般做法，使用生产者价格指数（PPI）或批发价格指数（WPI）来描述贸易品的价格变化，使用消费者价格指数（CPI）来描述非贸易品的价格走势。具体地，利用主要贸易伙伴的CPI 与PPI（或WPI）的比率，根据各自所占的权重，采用几何加权平均计算出“世界”
的非贸易品与贸易品的相对价格，然后用这一比率去除我国的CPI 与WPI 的比率，用对数表示为：
7
1
ln( ) ln( / ) ln( / ) i i i
i
tnt TNT CPI WPI w CPI WPI
=
= = -‡”
净对外资产（NFA）：把净对外资产作为实质汇率的一个决定因素，在开放经济宏观经济
学中有长久的历史，它是基于实质汇率决定的国际收支模型的考虑11。同样，在实际中，我
们得不到净对外资产的时间序列。作为替代，遵从文献的作法，利用积累的经常帐户
（accumulated current account）占GDP 的比率来模拟净对外资产的路径。然而，在我国
没有经常帐户余额的季度数据，本文的作法是用1990 年底外汇储备的余额作为1991 年初净
对外资产的替代，此后，利用每个季度的贸易盈余（出口减进口）作累计加法，得到累积的经常帐户的替代序列，最后使用这一序列与GDP 的比率，得到净对外资产NFA 的模拟数据,记这一比率为nfa。
贸易政策（TRADE）：贸易政策是影响实质汇率的一个重要因素，有大量文献发现发展中
国家贸易自由化的过程伴随着本国货币的贬值12。1990 年代以来我国贸易自由化的步伐不断加快，外贸体制改革，经常项目可兑换以及加入WTO 使我国在关税降低和非关税壁垒撤废等方面取得了实质进展，这些贸易政策的变化必然会对人民币实质汇率产生影响。然而很难到一个变量能够全面、有效地模拟贸易政策的影响。文献通常的做法是，利用进出口总额占GDP 的比率来描述一个国家的开放政策，并用该比例（即开放度）来模拟贸易政策对汇率的影响（如参见Elbadawi(1994)和Zhang（2001））。因为给定其它条件，越是自由化的贸易体制，贸易量也越大。在本文中，我们遵从这一做法。记这一比率为trade。
样本区间为1991 年1 季度至2004 年3 季度，数据为季度数据。所有的价格指数（包括
10 该效应是指贸易品与非贸易品部门生产率增长速度的差异会导致实质汇率的变化。其出发点是假设一国
的技术进步集中在贸易品部门，而非贸易品部门的技术创新则相对滞后。当前者的生产率提高时，其劳动
的边际产品增多，导致工资上升。如假设劳动力在部门之间自由流动，贸易品部门的工资上升会导致非贸
易品部门工资的上升，这样，整个非贸易品部门的成本上升，其结果是非贸易品价格的上升。而非贸易品
价格的上升又导致了国内总体价格水平的上升，从而带来本币的实质升值。
11 连续的经常项目赤字会导致该国净对外负债的增加，这需要用将来的贸易盈余来偿还。而实质汇率的贬
值将有助于产生这种贸易盈余，这意味着净对外资产余额的恶化导致中长期实质汇率的贬值。相反，本国
净对外资产余额的增加，会促进本国货币中长期的实质升值。基于跨时交易的实质汇率决定模型也将净对
外资产余额作为决定均衡实质汇率的重要基本面变量（如Obstfeld and Rogoff，1996）。
12 一般而言，封闭的落后国家能够动用的外汇储备非常有限，为了购买它们所急需的先进技术和关键设备，
不得不人为的制定一个较高的汇率，并且实行严格的贸易限制，以压制国内对国外普通商品的进口需求。
但在贸易自由化开始以后，这种过高的汇率便不再能够维系下去——外部平衡要求本国货币贬值到一个较
低的水平。
9
汇率）以1991 年1 季度为100，之后以此为基计算。国外及香港台湾地区的相关数据来源
于其政府官方统计网站； 我国的相关数据来源于高校财经数据库网站
（www.bjinfobank.com），贸易条件的季度数据来源于宋国青教授。图1-5 给出了上述变量
的时间序列图形。
图1 图2
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
1992 1994 1996 1998 2000 2002 2004
reer
4.52
4.56
4.60
4.64
4.68
4.72
1992 1994 1996 1998 2000 2002 200
tnt
图5
.2
.3
.4
.5
.6
.7
.8
1992 1994 1996 1998 2000 2002 2004
trade
图3 图4
.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1992 1994 1996 1998 2000 2002 2004
nfa
4.40
4.45
4.50
4.55
4.60
4.65
4.70
4.75
1992 1994 1996 1998 2000 2002 200
tot
10
[ 编辑 DadLU 在 08-04-23 05:06 ]
三、检验和估计13
在运用协整方法来确定实质汇率与经济基本面变量之间的长期均衡关系之前，需要先对
各个时间序列的稳态性进行分析，即运用单位根检验来判断数据的随机性质，然后建立一个
向量自回归模型（vector autoregression, 简称为VAR），最后利用协整方法确定各变量在
长期内的均衡表达式。
（一）、单位根检验
我们利用Enders（1995）给出的步骤对各变量进行单位根检验，选择适当的滞后阶数使
回归残差尽可能的接近白噪声。结果如表-1：
表-1 单位根检验（1991：q1－2004：q3）
截距时间趋势滞后阶数ADF 统计量10%关键值
reer 有无4 -2.231161 -2.5997
nfa 有有4 -3.076352 -3.1828
tnt 有有2 -2.256945 -3.1804
tot 有无0 -2.469564 -2.5970
trade 有有4 -1.095760 -3.1828
截距时间趋势滞后阶数ADF 统计量1%或5%关键值
D（reer） 无无4 -2.557282 -1.9480
D（nfa） 无无1 -2.537237 -1.9474
D（tnt） 无无1 -3.464298 -2.6100*
D（tot） 无无3 -3.343794 -2.6120*
D（trade） 无无3 -2.658581 -2.6120*
* 表示在1%的显著性水平下拒绝单位根假设；D 表示一阶差分。
从表-1 可以看出，各变量序列都存在单位根，而它们的一阶差分都在1%的显著水平下
拒绝了单位根假设，从而各变量都是I（1）序列。这为后面的协整检验提供了良好的基础。
（二）、VAR（2）估计
协整检验的前一步工作是估计一个不受限制的VAR （ unrestricted vector
autoregression）模型。鉴于样本长度的限制，本文选滞后阶数p=2，即估计VAR（2）。虽
然使用的是季度数据，p 取4 并不能带来更好的结果，相反，模型拟合的精度在下降（其主
要原因可能是样本点的限制）。考虑到季节性的影响，在模型中加入了截距项和三个集中化
的季节虚拟变量（censored seasonal dummies）14。表-2 给出了对VAR（2）系统的诊断结
果。
13 我们运用的计量软件是EViews 4.0。
14 如果加入的是0-1 型虚拟变量，则会改变VAR 中序列的均值和趋势；如引进集中化的虚拟变量，则只会
改变均值，对序列的趋势没有影响，详见Johansen（1995）。
11
表-2 VAR（2）模型的评估诊断（Model Evaluation Diagnostics）
Multivariate Diagnostic Test
Autocorrelation LM Test LM（1） LM（4） LM（8） LM（12）
36.23631 36.29299 33.58456 23.02931
（0.0681）* （0.0673） （0.1171） （0.5759）
Heteroskedasticity Test 2 (345) ÷ =377.1188 （0.1128）
Jarque-Bera normal Test 2 (10) ÷ =17.12299 （0.0717）
* 括号里的值表示p 值。
由表-2 所给出的评估诊断不难发现：异方差检验、8 阶和12 阶自相关检验都能通过，
但1 阶自相关检验以及正态性检验的p 值较小，表明残差序列出现1 阶自相关和非正态性的
可能性较大，这种风险可能会给我们后续的估计结果带来负面影响。通过分析数据我们发现，
实质有效汇率在1993 年第2 季度降到了最低点，并且在这一时点的前后分别出现了迅速的
下降和上升的过程。而在1998 年第1 季度，人民币的实质有效汇率上升到历史的最高水平，
之后开始下降。这表明可能出现了结构性的变化（structural breaks），如在我国汇率并轨
的前夕和东亚货币危机发生的时期，汇率的时间序列数据出现了结构性的变迁。为了反映这
一个问题，我们在上述VAR（2）模型中的1993 第2 季度和1998 年第1 季度加入两个0-1
型的虚拟变量，并重新估计。结果表明VAR（2）模型改进了很多。只有1 阶的自回归检验
的p 值较低，但在通常的显著性水平下能够拒绝1 阶自相关的假设。其它的自回归检验、White
异方差检验及Jarque-Bera 正态性检验都能较好地满足进一步计量分析的需要（参见文末的
附表）。不过根据两个模型得出的协整方程系数及其标准差的估计都非常接近，计算出的均
衡实质有效汇率几乎完全一样，所以我们下面仅报告对上述VAR（2）模型进行协整检验和
估计的结果。
[ 编辑 DadLU 在 08-04-23 05:08 ]
（三）、协整检验及其经济解释
本文采用Johansen 最大似然估计法进行协整检验，为此首先需要确定协整的具体形式15，
如截距项是否限制在协整空间里，协整变量是否采用具有趋势的形式。本文在比较了几种结
果之后，确定了最终的检验形式：协整变量具有线性趋势并且截距项限制在协整空间里。表
-3 给出了对上述VAR（2）模型的检验结果。从该结果可以看出，无论是迹统计量，还是最
大特征值统计量，都表明存在着一个协整关系。
表-3 VAR（2）的协整检验结果
协整秩H0 迹统计量5%临界值1%临界值
r=0** 78.44699 68.52 76.07
R 小于/等于 1 40.-9-8607 47.21 54.46
r 小于/等于 2 15.-3-1173 29.68 35.65
r 小于/等于 3 6.542232 15.41 20.04
r 小于/等于 4 0.706783 3.76 6.65
协整秩H0 最大特征值统计量5%临界值1%临界值
15 Eviews 4.0 提供了五种备选的形式，其中第一种和第五种很少使用。
12
r =0* 37.46091 33.46 38.77
r 小于/等于 1 25.67434 27.07 32.24
r 小于/等于 2 8.769503 20.97 25.52
r 小于/等于 3 5.835448 14.07 18.63
r 小于/等于 4 0.706783 3.76 6.65
*、*分别表示在5%、1%的显著性水平下拒绝原假设H0
表-4 给出了协整向量系数的估计值及其统计显著性水平。其中协整向量系数是以实质有
效汇率为基准，进行标准化后得到的结果，因而reer 前面的系数为1。调整系数使得我们
可以对汇率的调整过程有更加清楚的认识。从表-4 不难看出，所有的调整系数都带有正确
的符号，即所有的变量都趋向于自我稳定（self-stabilizing）。比如，当前一期的误差修
正项为正值时，即汇率高估，由于实质汇率的调整系数为-0.1032，在接下来的每个季度，
误差修正项都会以10.32%的比例减少汇率的高估。另外，从reer 的调整系数-0.1032 的大
小来看，半衰期大约为6 到7 个季度，与购买力平价（PPP）的结果比起来，汇率调整的速
度要快的多，后者表明半衰期通常为3 到5 年（Rogoff, 1996）。
表-4 VAR（2）模型的协整向量系数及调整系数的估计
reer nfa tnt tot trade c
1.000000 -0.498255 -2.220092 1.058739 0.775595 0.539198
协整向量
系数
(0.09697)
(0.72070) (0.41083) (0.23141)
调整系数-0. 103230 0.120209 0.025884 -0.101702 -0.197724
(0.05865) (0.03265) (0.01597) (0.07088) (0.09534)
*括号中的值表示（渐近）标准差。
四、计量结果的分析
（一）、人民币均衡实质汇率的决定因素分析
我们的计量结果给出的协整方程为：
= 0.498255 2.220092 1.058739 0.775595 0.539198 reer nfa tnt tot trade + - - - （6）
所有参数的估计值都显著的不为零，表明净对外资产占GDP 的比率、非贸易品-贸易品的相
对价格比、相对贸易条件、以及对外贸易政策都是人民币均衡实质汇率的重要长期决定因素。
其中，净对外资产占GDP 的比率及非贸易品-贸易品的相对价格比与均衡实质汇率成正向关
系，这与理论上的预期符合。净对外资产占GDP 的比率每增加1 个百分点会引起均衡实质汇
率升值0.498%；非贸易品-贸易品的相对价格比的影响则更为显著，其每增长1%将引起均衡
实质汇率升值2.22%。相对贸易条件与均衡实质汇率成反向关系，这说明贸易条件改善（恶
化）所产生的替代效应大于它所产生的收入效应（参见注7）。对外贸易政策变量对均衡实
质汇率具有负面的作用，这也与对发展中国家的许多研究文献相符，如Edwards（1994）、
Zhang（2001）也得出了类似的结论。理论上，发展中国家在对外开放的过程中，需要大量
进口一些关键的设备和技术，以及国民对国外产品的大量需求，在其它条件一定的前提下这
13
会使得开放的过程中伴随着均衡实质汇率的贬值。
（二）、人民币汇率失调的状况及其原因
把基本面变量（nfa，tot，tnt，trade）的实际值代入协整方程（6）就可以得到人民
币的当前均衡实质汇率。图-6 给出了人民币的当前均衡实质汇率（beer-current）和实际
实质汇率（reer）时序图。这两个汇率之间的差距表明了人民币汇率的当前失调程度。
图-6 人民币的当前均衡实质汇率和实际实质汇率
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
1992 1994 1996 1998 2000 2002 2004
reer beer_current
为了更加清楚地描述人民币实质汇率的失调程度，我们使用下述公式得出百分比表示的
[ 编辑 DadLU 在 08-04-23 05:34 ]
失调程度并把计算的结果放在图-7 中：16
100% × 实际实质汇率－当前均衡实质汇率
人民币汇率的当前失调＝
当前均衡实质汇率
图-7 人民币实质汇率的当前失调（百分比）
16 在计算百分比表示的汇率失调时，使用的是非对数化的均衡汇率，因此图-7 显示的失调程度要比图-6 大
一些。
14
-.3
-.2
-.1
.0
.1
.2
.3
1992 1994 1996 1998 2000 2002 2004
mis_current (%)
人民币汇率的当前失调虽然提供了关于人民币汇率失调便利和直接的信息，但是由于它
在计算过程中利用的是经济基本面的当前值，这些当前值包含了商业周期的影响。为了反映
均衡实质汇率决定中基本面变量持久性的而非一时性的影响，我们使用Hodrick-Prescott
（1980）滤波来提取基本面变量的长期均衡值17，并将它们代入协整方程（6）中，得到长期
（permanent）均衡汇率值，并把计算出来的长期均衡实质汇率值（beer-permanent）放在
图-8 中。类似于汇率当前失调的计算，我们利用下式来得出百分比表示的人民币汇率的长
期失调，并把计算结果放在图-9 中。
100% × 实际实质汇率－长期均衡实质汇率
人民币汇率的长期失调＝
长期均衡实质汇率
图-8 人民币的长期均衡实质汇率与实际实质汇率
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
1992 1994 1996 1998 2000 2002 2004
reer beer_permanent
17 还有很多其他的方法，如Beveridge-Nelson 法和移动平均法等，可以用来提取时间序列的长期均衡值。
15
图-9 人民币实质汇率的长期失调（百分比）
-.3
-.2
-.1
.0
.1
.2
.3
1992 1994 1996 1998 2000 2002 2004
mis_permanent (%)
把图-7 与图-9 进行比较可以看出，两者反映的人民币汇率失调的情况基本上是一致的。
不过，以基本面的长期值（或者说稳态值）决定的均衡实质汇率来衡量，人民币实质汇率的
失调情况表现得更为稳定。由于长期均衡实质汇率值（beer-permanent）能够更好地反映持
久性的（或者说稳态的）实质汇率与经济基本面变量之间的关系，所以我们以下的分析将集
中于对图-8 和图-9 的讨论。从图-8 和图-9 可以看出，人民币汇率失调的情况大体可以分为
三个时期来讨论：（1）1992 年2 季度–1994 年4 季度；（2）1995 年1 季度—1999 年2 季度；
以及（3）1999 年3 季度以后的时期。
在1992 年2 季度–1994 年4 季度期间人民币实质汇率表现为低估状态，其中以1993 年
2、3 季度最为严重，其低估程度分别达到了24.5%和22.5%。这一时期人民币实质汇率低估
的主要原因是“南巡讲话”后一轮高过一轮的投资热潮。过热的投资，需要进口大量先进的
技术和关键的设备，从而需要大量的外汇，这导致外汇调剂市场上人民币价格的走低（美元
价格的走高）18。1993 年外汇市场上人民币兑美元的调剂价由年初的5.7 元降至2 月的8.20
元，而5 月份更是暴跌至11.20 元。由于1990 年代初期外汇调剂市场的快速发展，人民币
与外币之间的交易越来越多地使用外汇调剂市场的汇率，调剂市场的汇率在人民币实质有效
汇率计算中所占的权重也越来越大（经由外汇调剂市场交易的比重高达80%）。虽然同期物
价上涨幅度也较大，但仍不能抵消调剂市场名义汇率贬值对实际实质汇率的影响。由于同期
均衡实质汇率的变化幅度相对较小，所以人民币实质汇率呈现为明显的低估状态。不过，1994
年高达24.2%的通货膨胀率还是使该年的低估程度趋于减小19。
由图-8 可以看出，从1994 年4 季度起人民币均衡实质汇率处于不断升值的状态，其背后的主要驱动力量是我国贸易品部门（制造业）劳动生产率的快速上升和经常项目盈余导致的净对外资产余额的不断增长（参见时序图）。同期人民币实际实质汇率则呈现出先快速升值后有所贬值和维持平稳的态势。人民币实际实质汇率的这种变化主要是受外部冲击，特别是1997-98 东亚货币危机和2002 年以来美元的大幅贬值的影响；人民币汇率政策（1994 年18 事实上，从1992 年第4 季度到1994 年第1 季度我国连续6 个季度出现贸易赤字（累积达到141.2 亿美元，而1992 年第4 季度的外汇储备仅为194.43 亿美元）。
19 1994 年元旦人民币汇率并轨，官方人民币名义汇率大幅贬值。但其对人民币实质汇率的影响极为有限。
一是按官方汇率交易的交易量已经很小，二是汇率并轨只是将官方汇率靠向当时的调剂市场汇率，而后者当时较稳定（还略有升值）。
16人民币汇率并轨以后，人民币名义汇率就基本保持不变，1997 年以后人民币更是钉住了美元）也是不可忽略的重要影响因素。结果，1995 年以来人民币实际实质汇率偏离均衡实质汇率的失调情况就表现为先是高估后为低估的状态。具体而言：
在1995 年1 季度—1999 年2 季度期间人民币实质汇率表现为高估。1995 年人民币均衡实质汇率基本保持不变，但该年17.2%的通货膨胀率（以消费者物价指数衡量）和人民币名义汇率稳定中略有升值推动人民币实际实质汇率上升从而导致了人民币实质汇率一定程度的高估20。在随后的三年（1996－1998 年），人民币实质汇率高估程度平均高达13.2%，在其间的个别季度高估甚至接近20%（这一点如果用当前失调衡量更为明显，严重的高估发生在1997 年1 季度—1998 年2 季度，1997 年1 季度人民币汇率高估程度达30%左右，见图7）。
这一期间人民币汇率高估既有外部冲击的原因，也人民币汇率政策的作用。众所周知，1997-1998 爆发了东亚货币危机，在此期间，除香港外我国主要贸易伙伴的货币对美元都出现了较大幅度的贬值：日元贬值了约30%，韩元的贬值超过50%（期间一度超过100%），新台币对美元的比价下降约25%，德国马克的贬值也在15%左右。而与此同时，我国政府采取了人民币钉住美元的政策，坚持人民币不贬值，这直接导致人民币名义有效汇率大幅升值。尽管在1997 年第2 季度，我国开始出现通货紧缩(以消费者价格指数的下降为衡量标准)，但由于紧缩的幅度相对较小，人民币实质有效汇率仍明显高于均衡实质汇率。1998 年底，随着东亚各国货币对外价值趋于稳定，我国持续的通货紧缩终于使得人民币汇率高估的情况得到了缓解。从1999 年3 季度往后的期间，人民币实质汇率重新转为低估的状态。这一时期人民币实质汇率失调表现为明显的结构性失调：由图2 可以看出这一期间非贸易品与贸易品的相对价格比保持了快速上升的势头，反映了我国制造业劳动生产率的持续上升，其结果人民币均衡实质汇率继续保持上升的趋势；然而由图1 可以看出，经济基本面的这种变化并没有反映
到人民币实际实质汇率的变化中：人民币实际实质汇率在这一期间大体维持不变甚至开始时有所下降！具体到失调的程度，1999 年3 季度－2002 年1 季度，人民币实质汇率低估的程度较轻，平均为4.24%。这一期间人民币实际实质汇率
[ 编辑 DadLU 在 08-04-23 05:35 ]
下降的原因主要有两个：一是东亚各国经济从危机中恢复，其货币对外价值趋于稳定，并对美元有较大幅度的升值21；二是我国通货紧缩的进一步发展。2002 年2 季度以后，人民币实质汇率低估的程度趋于严重，平均超过了10%并且有逐步扩大的趋势。此期间人民币实际实质汇率下降的原因主要是受到外部
冲击，即美元大幅贬值的影响22。钉住美元的汇率政策使人民币自动接纳了外部冲击，导致人民币的实际实质汇率也呈现大幅的下降，从而偏离经济基本面决定的人民币均衡实质汇率的轨迹。由于没有证据显示中国制造业劳动生产率的增长势头会放缓，而美国庞大的贸易赤字和财政赤字决定了美元在未来仍会呈贬值的趋势23，因此可以预见：如果我们继续实行人民币钉住美元的汇率政策人民币汇率低估的状态就会继续下去，并且汇率低估的程度会不断加深。20 同年，我国的主要贸易伙伴当中，只有香港的物价上涨了15%多一点，其它几国（或地区）都只有不到5%的上升。
21 如韩元对美元的比价从1998 年第1 季度的1606：1 下降到2000 年第4 季度的1167：1；日元对美元的比价在1998 年约为130.8：1，1999 年下降为113.8：1，2000 年进一步下降到107.8：1。
22 从欧元兑美元达到最低点的2000 年10 月20 日，到欧元兑美元达到最高点的2004 年11 月22 日，在4年多的时间里，美元对欧元贬值62.1%，同时，美元对英镑贬值34.2%，美元对日元贬值6.9%。用来衡量美元对一揽子货币的汇率变化程度的美元指数从119.07 下降到81.72，下降幅度达45.7%。
23 对美国对外经济政策有影响力的美国国际经济研究所所长伯格斯坦认为，为了重建美国可持续的外部平衡，美元名义有效汇率还需要进一步下降10-15%（Bergsten，2003）。
17
五、总结性评论
本文采用1991 年1 季度-2004 年3 季度的季度数据，运用BEER 模型估计了人民币均衡实质汇率和汇率失调程度。我们发现，（1）净对外资产占GDP 的比率、非贸易品-贸易品的相对价格比、贸易条件、以及对外贸易政策都是人民币均衡实质汇率的重要长期决定因素。其中，净对外资产占GDP 的比率及非贸易品-贸易品的相对价格比对均衡实质汇率具有正向作用，而贸易条件和对外贸易政策对均衡实质汇率具有反向作用；（2）从1994 年3 季度起人民币均衡实质汇率处于不断显著升值的状态，其背后的主要驱动力量是我国贸易品部门（制造业）劳动生产率的快速上升和经常项目盈余导致的净对外资产余额的不断增加。（3）上个世纪九十年代以来，人民币实际实质汇率在大部分时期偏离人民币均衡实质汇率轨迹，表现为人民币实质汇率的失调。其中，1992 年2 季度–1994 年4 季度为人民币汇率低估时期，并以1993 年2、3 季度最为严重；1995 年1 季度—1999 年2 季度为人民币汇率明显的高估时期，而较为严重的高估发生在1997 年1 季度—1998 年2 季度；1999 年3 季度往后的时期人民币汇率重新转为明显的低估，并且低估程度有进一步扩大的趋势。持续的汇率失调使中国经济付出了以资源配置的效率损失和国民福利的降低来衡量的高昂代价，例如，Tyers & Yang（2000）运用一个全球比较静态一般均衡模型就东亚货币危
机期间主要冲击对中国经济的影响，分固定汇率制度和浮动汇率制度分别进行了模拟实验,其结果表明如果东亚货币危机期间人民币贬值，中国本可以每年避免相当于GDP 百分之四的损失。值得指出的是：我们的分析表明，1997 年以来事实上的钉住美元的汇率政策是造成人民币汇率失调的一个主要的宏观政策因素。钉住美元的汇率政策拖延了经济对汇率失调的及时调整并使调整的代价过大：在人民币汇率明显低估（高估）的状况下，因为人民币名义汇率不能对这种失调压力做出反映，经济只有通过通货膨胀（通货紧缩）或其它内在机制进
行调整，后者的调整成本显然更大。因此，本研究的政策含义是，从应对人民币汇率失调的角度，一个更为灵活的人民币汇率制度将更有利于中国经济的健康发展。
参考文献：
Baffes J., Elbadawi I.A., O’Connell S.A. (1999), “Single–Equation Estimation of
the Equilibrium Real Exchange Rate”, In Hinkle, L.E. and Montiel, P.J. (ed.),
Exchange Rate Misalignment: Concepts and Measurement for Developing countries,
pp.405-465, A World Bank Research Publication, Oxford: Oxford University
Press.
Bergsten，Fred（2003）：The Correction of the Dollar and Foreign Interventionin the
Currency Markets ， Testimony Before the Committee on Small Business
UnitedStates House Representative Washington，DC.
Chinn, M. C. (1999)，“Productivity, government spending and the real exchange rate:
Evidence for OECD countries”, in R. MacDonald and J. L Stein (ed) Equilibrium
ExchangeRates, Kluwer Academic Publisher, UK, 163-190。
Clark, P. B. and R. MacDonald (1999), “Exchange Rates and Economic Fundamentals:
A Methodological Comparison of BEERs and FEERs”, in MacDonald, R, and Stein,
J, (eds) Equilibrium Exchange Rates, Kluwer Academic Publishers.
Cottani, J.A., D.F. Cavallo and M.S. Khan (1990), “Real Exchange Rate Behavior and
18
Economic Performance in LDCs”, Economic Development and Cultural Change, (39),
61-76.
Clostermann, J. and B. Schnatz (2000) “The determinants of the euro-dollar exchange
rate Synthetic fundamentals and a non-existing currency”, Applied Economics
Quarterly, 46, 3, 274-302.
Edwards, S.(1988), Exchange Rate Misalignment in Developing Countries. Washington,
DC: The World Bank.
Edwards, S. (1994). “Real and Monetary Determinants of Real Exchange Rate
Behavior:Theory and Evidence from Developing Countries”, In J. Williamson
(ed.), Estimating Equilibrium Exchange Rates. Washington D.C.: Institute of
International Economics.
Edwards, S and M. Savastano (1999), “Exchange Rates in Emerging Economies: What
Do We Know? What Do We Need To Know?”, NBER Working Paper No. 7228.
Elbadawi, Ibrahim A. (1994), “Estimating Long-Run Equilibrium Real Exchange
Rate.” In John Williamson (Ed.), Estimating Equilibrium Exchange Rates,
Washington, DC: Institute for International Economics
Enders, W. (1995), Applied Econometric Time Series, John Wiley & Sons, Inc.
Faruqee, H. (1995), “Long-Run Determinants of the real Exchange rate: A Stock-Flow
Perspective”, IMF Staff Papers, Vol. 42, pp. 80-107.
Maeso–Fernandez, F.; Osbat, C.; Schnatz, B （2002），“Determinants of the Euro Real
Effective Exchange Rate: A BEER/PEER Approach ”， Australian Economic Papers,
Dec2002, Vol. 41 Issue 4, p437, 25p
Hodrick, R and E. Prescott (1980), ‘Post-war US business cycles: an empirical
investigation’, Unpublished manuscript, Carnegie Mellon University.
Johansen,S. (1995), Likelihood-Based Inference in Cointegrated Vector
Autoregressive Models, Oxford: Oxford University Press.
Montiel, P. J.（1999a） “The Long-Run Equilibrium Real Exchange Rate: Conceptual
Issues and Empirical Research.” In L. Hinkle and P. J. Montiel, Eds., Exchange
Rate Misalignment: Concepts and Measurement for Developing Countries., A World
Bank Research Publication, Oxford: Oxford Univ. Press, pp. 219–263
Montiel, Peter J. (1999b), “Determinants of the Long-Run Equilibrium Real Exchange
Rate: An Analytical Model”, In Hinkle L.E. and Montiel P. J. (Ed.), Exchange
Rate Misalignment: Concepts and Measurement for Developing countries, A World
Bank Research Publication, Oxford: Oxford University Press, pp. 264-290.
Obstfeld, M. and Rogoff,K.(1995) “Exchange Rate Dynamics Redux” Journal of
Political Economy 103:624-60.
Obstfeld，M. and Rogoff，K. （1996）, Foundations of International Macroeconomics，
MIT press，Cambridge MA，pp199-268。
Rapach, D. (2001), “Monetary Shocks and Real Exchange Rate Hysteresis: Evidence from
the G-7 Countries”, Review of International Economics, 9(2), 356-371.
Rogoff, R. (1996), “The Purchasing Power Parity Puzzle”, The Journal of Economic
Literature,Vol.34, pp.647-668
Tyers, R. and Y. Yang, （2000）,“Weathering the Crisis: The Role of China”，
Pacific Economic Papers 308，Asia-Pacific School of Economics and Management,
19
Australian National University。
Williamson, J. (1985), The Exchange Rate System, 2nd ed, Washington D.C.: Institute
of International Economics.
Williamson, J. (1994), “Estimates of FEERS”, In J. Williamson (ed.), Estimating
Equilibrium Exchange Rates. Washington D.C.: Institute of International
Economics.
Wren-Lewis, S. (1992), “On the Analytical Foundations of the Fundamental
Equilibrium Exchange Rate”, In C. P. Hargreaves (ed.), Macroeconomic Modeling
of the Long Run, Edward Elgar.
Zhichao, Zhang, (2001), “Real Exchange Rate Misalignment in china: An Empirical
Investigation”, Journal of Comparative Economics, Vol. 29, pp.80-94.
张晓朴（1999）,“人民币均衡汇率理论与模型”，《经济研究》，年第12 期。
林伯强（2002）,“人民币均衡实际汇率的估计与实际汇率错位的测算”，《经济研究》，
第12 期。
张斌（2003），“人民币均衡汇率：简约一般均衡下的单方程模型研究”，《世界经济》，第11
期。
Renminbi Equilibrium Exchange Rate and China’s Exchange Rate
Misalignment: 1991-2004
Jianhuai Shi
(China Center for Economic Research, Peking University)
Haifeng Yu
(Asset Management Department, Hongta Securities Company Ltd.)
Abstract: This paper employs the so called behavioral equilibrium exchange rate model to estimate the
equilibrium real exchange rate of Renminbi and the exchange rate misalignment in China, which covers the period
from 1991q1 to 2004q3. The main findings of the paper are that (1) since 1994q4, Renminbi equilibrium
exchange rate has exhibited a steady appreciation, the main moving forces behind that are the fast productivity
growth within the manufacture sector and the rapidly enlarged net foreign asset position; and (2) that Renminbi
real exchange rate has been misaligned during most of the sample period, particularly, undervaluation occurred in
1992q2—1994q4 period, overvaluation occurred in 1995q1—1999q2, and undervaluation has occurred again since
1999q3, which also has an enlarging trend since 2002q2. The paper also analyses the economic reasons and
policy implications of the findings.
Key words: Renminbi, equilibrium real exchange rate, exchange rate misalignment, behavioral equilibrium
exchange rate model
20
附表：加入0-1虚拟变量的VAR（2）模型的检验和估计结果
表-6 加入0-1 虚拟变量的VAR（2）模型的协整检验结果
协整秩H0 迹统计量5%临界值1%临界值
r=0** 81.38691 68.52 76.07
r小于或等于1 44.5-4488 47.21 54.46
r小于或等于2 18.7-8150 29.68 35.65
r小于或等于3 7.0-53751 15.41 20.04
r小于或等于 4 1.2-87372 3.76 6.65
协整秩H0 最大特征值统计量5%临界值1%临界值
r=0* 36.84203 33.46 38.77
r
未完，但是苹果的码弄不上了
[ 编辑 DadLU 在 08-04-23 05:38 ]
谢谢大家都这么自觉
我突然发现我再占1000 个位都说不完
不过等我再占10个大家再顶 
气呀！！！
支持中国！！！！！！
无条件支持！！！！！！！！！！
中国！
我爱你！！！
我们要理智的爱国！！！
[ 编辑 DadLU 在 08-04-23 05:43 ]
原来愚昧也是不爱国的表现 
我要好好研究 + 实践！！