Understanding Chinese GDP [1st ed. 2019] 978-981-32-9732-6, 978-981-32-9733-3

Table of contents :
Front Matter ....Pages i-xxi
One Dropped Pebble Creates a Thousand Ripples (Xuguang Song)....Pages 1-15
Re-examining Some Problems in the Methodology of International Economic Comparison (Xuguang Song)....Pages 17-36
Limitations of the Exchange Rate Method (Xuguang Song)....Pages 37-53
Viewing and Applying the PPP Correctly (Xuguang Song)....Pages 55-95
Understanding PPP Through Examples (Xuguang Song)....Pages 97-106
Calculating China’s GDP (Xuguang Song)....Pages 107-132
Confusion in the Service Industry Data (Xuguang Song)....Pages 133-167
Similarities and Differences Between GDP and GNI (Xuguang Song)....Pages 169-199
National Power and Processing Depth Coefficient (Xuguang Song)....Pages 201-232
Comparisons of People’s Standard of Living (Xuguang Song)....Pages 233-247
Rankings of Countries in Terms of Energy, Per Capita Arable Land, Water Resources, and Expenditures on Medical and Health Care (Xuguang Song)....Pages 249-254
Changes in the GDP Rankings (Xuguang Song)....Pages 255-287
Economic Scale, the United Nations Membership Dues, and Shares of the World Bank (Xuguang Song)....Pages 289-303
ICP Shock (Xuguang Song)....Pages 305-316
Back Matter ....Pages 317-355

Citation preview

Understanding Chinese GDP x ugua ng song

Understanding Chinese GDP

Xuguang Song

Understanding Chinese GDP

Xuguang Song Beijing Normal University Beijing, China Translated by Jin Chen

Funded by the Chinese Fund for the Humanities and Social Sciences

ISBN 978-981-32-9732-6 ISBN 978-981-32-9733-3  (eBook) https://doi.org/10.1007/978-981-32-9733-3 Jointly published with Peking University Press The print edition is not for sale in Mainland of China. Customers from Mainland of China please order the print book from: Peking University Press. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2019 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publishers, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publishers nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publishers remain neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Palgrave Macmillan imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

Does China’s GDP rank No. 1 or No. 2 in the world? After the World Bank released the results of the International Comparison Program (ICP) on April 30, 2014, this question has become a hot topic and has caught the attention of many people. When studying the ranking of China’s GDP in the world, we think of a poem written by Su Dongpo, a poet in the Song Dynasty: It’s like a range when you look at the mountain from the front. But it’s like a peak when you look at it sideways. The mountain shows its different features In different levels near and far. You don’t know the real Lushan Mountain, Because you are in the mountain yourself.

People who say Lushan Mountain is a range or a peak are all correct, because they look at the mountain from different perspectives. Similarly, those who say China’s GDP ranks No. 1 or No. 2 in the world are all right, because they focus on different aspects. It is more complex to measure a country’s GDP than observing a mountain. Some people pay attention to the economic aggregate, some focus on the per capita index, and others care about the living standards, so their ­conclusions are totally different. v

vi  

PREFACE

Just as the report of ICP of the World Bank says, according to the PPP method, China’s GDP has exceeded that of the United States to rank No. 1 in the world; however, according to the exchange rate method, it is only half of that of the United States. In terms of the material output, China’s GDP exceeded that of the United States as early as 2002; however, in terms of the standard of living, in 2013, the US per capita GDP was 4.46 times (by the PPP method) or 7.80 times (by the exchange rate method) that of China and the US per capita consumption was 5.32 times that of China. The gap between the two countries is still quite big. ICP led by the World Bank is committed to research on the PPP method. It has achieved great progress. However, due to the inherent contradiction between the theory and the statistical method, the PPP method still needs to be further improved. At present, the result obtained by the PPP method can at best provide a specific angle for observation, which still has a long way to go before people can use it to develop relevant policies. Since its founding in 2011, the National Accounting Institute of Beijing Normal University has undertaken a series of national accounting research tasks and made some achievements, in which the National Accounting Research Report 2013 and the National Accounting Research Report 2014 have won a universal appraisal from academic and business communities and the government policy research department. Since the issues about China’s GDP have received widespread attention at home and abroad, on the basis of previous studies, the colleagues in the National Accounting Institute worked together and tackled key problems to write this book in just a few months. Chapter 1 of the book is written by Song Xuguang and Xu Dianqing; Chapter 2 by Qiu Dong; Chapter 3 by Wang Yafei and Xu Dianqing; Chapter 4 by Chen Menggen, Wang Yafei, Hu Xuemei, Li Xin, Ding Mengmeng, and Xu Dianqing; Chapter 5 by Wang Yafei, Li Xin, and Xu Dianqing; Chapter 6 by Wang Yafei; Chapter 7 by Xi Wei, Ding Mengmeng, Li Xin, and Xu Dianqing; Chapter 8 by Lv Guangming, Xu Man, and Jia Shuaishuai; Chapter 9 by Li Xin, Du Yonghong, Ding Mengmeng, and Xu Dianqing; Chapter 10 by Ding Mengmeng and Xu Dianqing; Chapter 11 by Ding Mengmeng and Xu Dianqing; Chapter 12 by Ding Mengmeng, Li Xin, and Xu Dianqing; Chapter 13 by Li Xin, Hu Xuemei, and Xu Dianqing; and Chapter 14 by Song Xuguang, Chen Menggen, Wang Yafei, and Xu Dianqing. In the process

PREFACE  

vii

of writing, colleagues at the National Accounting Institute of Beijing Normal University learned from each other by exchanging views and collaborating with each other. Research assistants Wang Luyao and Yang Shuo and many graduate students also actively participated in data collection and organization. Collective effort is a sure warrant for the publication of this book. Hearty thanks are given to Liu Guoguang, Wu Jinglian, Mao Yushi, Zhang Shuguang, Zhang Weiying, Lu Feng, Yao Yang, Li Ling, Li Shi, Li Xiaoxi, Tang Renwu, Xie Ping, Tang Min, Zuo Xiaolei, Li Shantong, Zheng Yuxin, Zhao Haiying, Ren Ruoen, Cao Yuanzheng, Zuo Xuejin, and Shi Jinchuan who have provided guidance and help in the process of writing, and to Lin Junxiu, Hao Xiaonan, and Huang Weiting from Peking University Press who have made significant contributions to the publication of the book. We know clearly that it is impossible to give a perfect answer to this complicated problem due to our limited knowledge. There may be many omissions and erroneous opinions needing further correction and improvement in the book. Our purpose is just to stimulate public discussion on the topic. We sincerely welcome the comments and advices of all experts. Beijing, China November 2014

Xuguang Song

Contents

1

One 1.1 1.2 1.3 1.4 1.5

Dropped Pebble Creates a Thousand Ripples 1 Dispute Arising from a World Bank Report 1 Long-Standing Debate 3 Difficulties in Measuring Economic Scale 6 Two Sets of GDP Statistical Data 7 Different Statistical Conclusions Can Be Reached Through Different Observation Perspectives 10 1.6 The World Bank’s Two Sets of Rankings 12

2

Re-examining Some Problems in the Methodology of International Economic Comparison 17 2.1 Premise Setting of Comparison Homogeneity 17 2.2 Rethinking the ICP and Exchange Rate Methods 22 2.3 Spatial Structure and Its Measure Comparison 29 2.4 Problems that Need Further Research in International Economic Comparison and Other Suggestions 33

3

Limitations of the Exchange Rate Method 37 3.1 The Exchange Rate and PPP Methods 37 3.2 Inherent Contradictions of the Exchange Rate Method 39 3.3 Origin of Official GDP Data by the Exchange Rate Method 47 3.4 Premises of Using the Exchange Rate Method 52 ix

x 

CONTENTS

4

Viewing and Applying the PPP Correctly 55 4.1 Currency Purchasing Power and PPP 55 4.2 Origin and Progress of ICP 59 4.3 Methods of ICP 66 4.4 Basic Application of ICP 73 4.5 ICP in China 80 4.6 Challenges Faced by the ICP 83 4.7 Misconceptions About the PPP 88 References 94

5

Understanding PPP Through Examples 97 5.1 Simple Examples of GDP Calculation 97 5.2 Calculation Rules of the PPP Method 99 5.3 Problems of the PPP Method 100 5.4 Do Not Take the Experience of One Point and Spread It to the Entire Area or Take a Part for the Whole 103

6

Calculating China’s GDP 107 6.1 Two Types of Accounting Systems 107 6.2 Evolution of China’s National Accounting 111 6.3 Major Adjustments of the Official Statistical Data 115 6.4 The Sum of Regional GDPs Is Greater Than the National GDP 118 6.5 Reasons Why the Sum of Regional GDPs Is Greater Than the National GDP 128

7

Confusion in the Service Industry Data 133 7.1 Seriously Distorted Proportion of China’s Service Industry 133 7.2 Several Major Adjustments in China’s Service Industry Data 136 7.3 Empirical Analysis of the GDP Proportion of the Service Industry 142 7.4 Statistical System Transformation and the Remaining Issues 155 7.5 Reasons for the Loss of China’s Service Industry Data 156 7.6 Estimating GDP After Adjusting the Proportion of China’s Service Industry 162

CONTENTS  

xi

8

Similarities and Differences Between GDP and GNI 169 8.1 Gross Domestic Product (GDP) and Gross National Income (GNI) 169 8.2 Explanation of China’s GNI Accounting 172 8.3 Analysis of Differences Between China’s GDP and GNI 173 8.4 Quality Investigation of China’s GNI Data from the Perspective of Balance of Payments 178 8.5 Investigation on the Quality of China’s GNI Data from the Perspective of External Data Verification 183 8.6 Analysis of Factors Affecting the Differences Between GDP and GNI 189 8.7 Summary and Policy Suggestions 196 References 199

9

National Power and Processing Depth Coefficient 201 9.1 Reference Index Measuring Scale of Economy: National Power 201 9.2 Rankings of Total Material Output 207 9.3 Processing Depth Coefficient Method 218 References 232

10 Comparisons of People’s Standard of Living 233 10.1 Comparison of Per Capita GDPs According to the PPP Method 233 10.2 Comparison of Per Capita GDPs According to the Exchange Rate Method 241 11 Rankings of Countries in Terms of Energy, Per Capita Arable Land, Water Resources, and Expenditures on Medical and Health Care 249 11.1 Rankings in Terms of Energy 249 11.2 Rankings in Terms of Per Capita Arable Land 250 11.3 Rankings in Terms of Water Resources 251 11.4 Rankings According to Expenditure on Medical and Health Care 252

xii  

CONTENTS

12 Changes in the GDP Rankings 255 12.1 Different GDP Rankings from Different Perspectives 255 12.2 Changes in GDP Rankings According to the PPP Method 256 12.3 Changes in GDP Rankings According to the Exchange Rate Method 258 12.4 Changes in GDP Rankings According to Material Output 260 12.5 GDP Rankings According to Per Capita GDP 265 12.6 GDP Rankings According to Per Capita Consumption 268 12.7 GDP Rankings After Adjustments of the Proportion of the Service Industry 278 12.8 GDP Rankings According to Agricultural and Industrial Output 282 12.9 Has China’s Economic Size Exceeded that of the United States? 284 13 Economic Scale, the United Nations Membership Dues, and Shares of the World Bank 289 13.1 Rules for Paying UN Membership Dues 289 13.2 Evolution of China’s Payment of UN Membership Dues 291 13.3 Payments to International Financial Institutions and Voting Rights 293 14 ICP Shock 305 14.1 Opportunities and Challenges Brought About by the ICP 305 14.2 Disadvantages of the ICP 308 14.3 How to View the ICP Results 311 14.4 How to View the Data Provided by the PPP Method 314 Appendix A: G  DP Rankings by Country Using the PPP Method and the Exchange Rate Method 317 Appendix B: Design of the PPP Homogeneity Index 329 Appendix C: GDP Share of the Service Industry by Country 337 References 349

List of Figures

Fig. 3.1 Fig. 4.1

Fig. 4.2 Fig. 6.1 Fig. 6.2 Fig. 6.3 Fig. 6.4

Fig. 6.5

Japanese GDP growth rates at different currencies between 1960 and 2011 (Data Source WDI, 2012 ed. World Bank) 44 Trends of China’s PPP and the average exchange rate of the RMB 1990–2015 (Source Exchange rates from China Statistical Yearbook, and PPP data from http://databank. worldbank.org/data/views/variableselection/selectvariables. aspx?source=wo) 90 Intersection of commodities and services between two countries 92 Trends of national GDP and the sum of regional GDPs between 1990 and 2012 (Source China Statistical Yearbook for respective years, National Bureau of Statistics of China) 119 Differences between sum of regional GDPs and national GDP from 1990 to 2012 (Source China Statistical Yearbook, National Bureau of Statistics of China) 120 Differences between sum of regional GDPs and national GDP by industry between 1990 and 2012 (Source China Statistical Yearbook, National Bureau of Statistics of China) 121 1990–2012 differences between sum of regional GPDs and national GDP in the secondary industry by sector (Source China Statistical Yearbook, National Bureau of Statistics of China) 124 Trends of national GDP between 2000 and 2012 by the expenditure method (Source China Statistical Yearbook, National Bureau of Statistics of China) 125

xiii

xiv  

LIST OF FIGURES

Fig. 6.6

Fig. 6.7

Fig. 7.1 Fig. 7.2 Fig. 7.3 Fig. 7.4 Fig. 7.5 Fig. 7.6 Fig. 7.7

Fig. 7.8 Fig. 8.1 Fig. 8.2 Fig. 8.3 Fig. 8.4 Fig. 9.1

Differences between the sum of regional GDPs and the national GDP between 2000 and 2012 using the expenditure method (Source China Statistical Yearbook, National Bureau of Statistics of China) 126 Differences between the sum of regional GDPs and the national GDP by constituent items from 2000 to 2012 using the expenditure method (Source China Statistical Yearbook, National Bureau of Statistics of China) 127 Adjustment of China’s service industry as a percentage of GDP 146 Comparison between before and after GDP adjustment (PPP method) 149 GDP share of the service industry. a Before adjustment. b After adjustment 152 Forecast on GDP share of the service industry. a Before adjustment. b After adjustment 153 Forecast trend of GDP. a Before adjustment. b After adjustment 154 The GDP proportion of service industry and per capita GDP across countries worldwide in 2012 162 Reasonable share of service industry relative to per capita GDP in 2009. Note A0 represents per capita GDP of the least developed countries, A1 represents per capita GDP of low-income countries, A2 represents per capita GDP of mid-low-income countries, and A3 represents per capita GDP of mid-high-income countries (Data Source World Development Index [WDI], World Bank, 2011) 163 GDP after adjusting the proportion of service industry 167 Trends of differences between GNI and GDP in the current account balance, the employee compensation, and the investment income from 1982 to 2012 176 Creditor, debtor, and difference of investment income between 1982 and 2012 176 Trends of China’s GNI and GDP data from different data sources 186 Clustering analysis of China’s GNI and GDP indices from different sources 188 Relationship between processing depth coefficient and per capita GDP. Note The horizontal axis represents countries’ processing depth coefficient, and the vertical axis represents the level of per capita GDP (thousands of US dollars)

LIST OF FIGURES  

Fig. 12.1 Fig. 12.2 Fig. 12.3 Fig. 12.4 Fig. 12.5 Fig. 12.6

for sample countries. The convex curve in the figure represents the fitting line of sample points (Data Source World Development Indicators, 2012) GDP of selected countries between 1990 and 2013 using the PPP method GDP of selected countries between 1990 and 2013 according to the exchange rate method GDP excluding service industry for selected countries between 1990 and 2013 using the PPP method GDP excluding service industry for selected countries between 1990 and 2013 (Exchange Rate Method) (Data Source World Bank Database, 2014) Per capita GDP by PPP method Per capita GDP by the exchange rate method

xv

229 257 259 262 265 267 268

List of Tables

Table 1.1 Table 1.2 Table 1.3 Table 1.4 Table 1.5 Table 1.6 Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 4.1 Table 4.2 Table 4.3 Table 4.4

GDP estimated by the World Bank using the purchasing power parity method (Unit: 100 million USD) 2 Comparison of Sino-US economic scale calculated by American Professor Stephens (Unit: Trillion USD) 5 China’s GDP calculated according to the exchange rate method 8 Gaps of GDP between the two methods 9 Horizontal comparison of GDP measured in purchasing power parity terms (Unit: 100 million USD) 13 Horizontal comparison of GDP measured in purchasing power parity terms (Unit: 100 million USD) 14 Effects of exchange rate fluctuations on the estimates of Japanese GDP 43 Economic growth rates of China 46 GDP at current price for China, the United States, and Japan from 1960 to 2013 using the exchange rate method (Unit: billion USD) 49 GDP at constant price for China, the United States, and Japan from 1960 to 2013 using the exchange rate method (Unit: billion USD) 50 Number of participating countries in ICP 61 Distribution of product specifications in the eighth round of ICP 64 The eighth round ICP classification of expenditures 65 PPP conversion factor of China, GDP 81

xvii

xviii  

LIST OF TABLES

Table 4.5 Table 4.6 Table 4.7 Table 5.1 Table 6.1 Table 6.2 Table 6.3 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 7.6 Table 7.7 Table 7.8 Table 7.9 Table 7.10 Table 7.11 Table 7.12 Table 7.13 Table 7.14 Table 8.1 Table 8.2 Table 8.3

International rankings of China in terms of PPP and the relative price level Chinese GDP in LUC and in US dollar calculated by the PPP conversion factor Advantages and disadvantages of different PPP methods Case of purchasing power parity Differences between national GDP and sum of regional GDPs by industry 1990–2013 (Unit: 100 million CNY) Differences between national GPD and the aggregated regional GDPs within the secondary industry in 1990–2013 (Unit: 100 million CNY) 2000–2012 difference between sum of regional GDPs and national GDP by composition using the expenditure method (Unit: 100 million CNY) Average of the service industry’s share of GDP (group by revenue) (Unit: %) Average of the service Industry’s share of GDP by region (Unit: %) China’s GDP values 2004 GDP data adjustment (Unit: 100 million CNY) Descriptive statistics of variables Variable unit root test Regression analysis of service industry’s share of GDP and GDP per capita GDP proportion of china’s service industry (%) Comparison between before and after GDP adjustment (PPP method, current price, and international dollar) Robustness check for regression coefficients Unit root test Estimates of Chinese GDP in 2009 with different shares of the service industry GDP with adjusted proportions of the service industry in 2009 (Unit: 100 million USD) GDP after adjusting proportions of the service industry (PPP) (Unit: 100 million USD) Difference between GDP and GNI and difference in balance of current accounts from 1982 to 2012 (Unit: 100 million CNY) Analysis on GNI accounting and balance of international payments Deviations of the virtual exchange rates I from the real exchange rates prior to the exchange rate reform in 2005

81 82 86 98 122 123 126 134 135 140 141 143 144 145 147 148 150 151 164 165 166 174 179 181

LIST OF TABLES  

Table 8.4 Table 8.5 Table 8.6 Table 8.7 Table 8.8 Table 8.9 Table 8.10 Table 9.1 Table 9.2 Table 9.3 Table 9.4 Table 9.5 Table 9.6 Table 9.7 Table 9.8 Table 9.9 Table 10.1 Table 10.2 Table 10.3 Table 10.4 Table 10.5 Table 10.6 Table 11.1

xix

Fluctuations of the monthly average exchange rates for USD to CNY from 2006 to 2012 181 Three-way comparisons of GNI and GDP adjusted by exchange rate (Unit: 100 million CNY) 185 Mean and standard deviation of China’s GNI and GDP from different data sources (Unit: 100 million CNY) 187 Deviation of China’s GNI and GDP indices from common tendency 189 China’s FDI inflows from 1986 to 2012 190 Proportions of foreign-invested enterprises in total imports and exports from 1986 to 2012 (Currency unit: 100 million USD) 192 Import and export of China’s processing trade and its proportions from 1985 to 2012 (Currency unit: 100 million CNY) 195 Rankings of comprehensive national power by various institutions 204 Country rankings by national power in 2009 206 World ranking of GDP excluding service industry (PPP) (Unit: 100 million USD) 208 World rankings of GDP excluding service industry (exchange rate method) (Unit: 100 million USD) 213 Countries used for estimating GDP with the processing depth coefficient method 220 Consumption of industrial raw materials across countries in 2011 221 Major agricultural products by country in 2011 223 Estimated virtual output value and processing depth coefficients across countries in 2011 226 Horizontal comparisons with adjusted Chinese GDP 230 Per capita GDP of selected countries (PPP) (Unit: USD) 234 Per capita GDP of selected countries as compared to that of China (PPP) 234 Per capita GDP ranking estimated by purchasing power parity (Unit: dollar) 235 Per capita GDP across countries (exchange rate method) (Unit: dollar) 241 Per capita GDP of selected countries as compared to that of China (exchange rate method) 241 Ranking of per capita GDP estimated by exchange rate method (Unit: dollar) 242 Total energy production and per capita output by country 250

xx  

LIST OF TABLES

Table 11.2 Table 11.3 Table 11.4 Table 11.5 Table 12.1 Table 12.2 Table 12.3 Table 12.4 Table 12.5 Table 12.6 Table 12.7 Table 12.8 Table 12.9 Table 12.10

Table 12.11

Table 12.12 Table 12.13 Table 13.1 Table 13.2 Table 13.3 Table 13.4

Country rankings by per capita energy use in 2011 Country rankings by per capita arable land in 2012 Country rankings by renewable internal freshwater resources in 2013 Country ranking by health expenditure per capita in 2012 (PPP) Estimated national GDP of selected countries using the PPP method (Unit: 100 million USD) National GDP of selected countries as compared to that of China using the PPP method Estimated national GDP of selected countries using the exchange rate method (Unit: 100 million USD) National GDP of selected countries as compared to that of China using the exchange rate method Economic size after excluding service industry (PPP) (Unit: 100 million USD) Growth rate of material output for selected countries between 1991 and 2013 (PPP) (Unit: %) Economic size after excluding service industry (exchange rate method) (Unit: 100 million USD) Ranks of China and India in terms of per capita GDP by different methods Ranking of economies in terms of per capita consumption in 2011 Ranking of top five countries by GDP with the proportion of China’s industry service at the average level of low-income countries (PPP) (Unit: 100 million USD) Ranking of top five countries by GDP with the proportion of China’s industry service at the average level of lower-middle-income countries (PPP) (Unit: 100 million USD) World rankings of China by different indices China’s economy as compared to that of the United States and Japan Share and amount of UN Dues paid by Member States in 2013 China’s share of UN membership dues Countries with large changes in voting power before and after the reform in 2010 (Unit: %) Subscription for shares of IFC’s capital stock

251 252 253 254 256 256 259 260 261 263 264 266 269

279

281 283 285 290 292 295 296

LIST OF TABLES  

Table 13.5 Table A.1 Table A.2 Table C.1

Subscription for shares of IBRD’s stock capital and voting power of Member States (as of October 10, 2013) GDP rankings by country using the PPP method (Unit: 100 million USD) GDP rankings by country using the exchange rate method (Unit: 100 million USD) GDP share of the service industry by country (%)

xxi

298 318 323 338

CHAPTER 1

One Dropped Pebble Creates a Thousand Ripples

1.1   Dispute Arising from a World Bank Report On April 30, 2014, the World Bank published the report Purchasing Power Parities and Real Expenditures of World Economies: Summary of Results and Findings of the 2011 International Comparison Program, a summary of the preliminary results of the Eighth Round International Comparison Program (ICP 2011). According to this report, China’s purchasing power parity (PPP) was 3.506 in 2011. Based on this statistic, China’s GDP in 2011 was 47.16 trillion Chinese yuan (CNY) or an equivalent of 13.46 trillion US dollars (USD) after conversion, accounting for 14.9% of the global share. The GDP in the United States was USD 15.53 trillion over the same period, accounting for 17.1% of the global share. China’s economic scale reached 87.0% of that of the United States in 2011, indicating that China’s economic scale was already approaching the United States then. Further data from the World Bank shows that China’s GDP in 2013 was USD 16.16 trillion, while that of the United States was USD 16.8 trillion (see Table 1.1); these numbers clearly indicate that only a small difference could be observed between two countries’ economic scales. In 2013, China’s GDP growth rate was 7.7%, while that of the United States was 1.6%.1 China’s economic

1 The

GDP growth rate of the United States in 2013 is from www.cia.gov.

© The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_1

1

2  X. SONG Table 1.1 GDP estimated by the World Bank using the purchasing power parity method (Unit: 100 million USD)

Year

China

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

11,427 12,260 13,691 16,012 18,637 21,514 24,301 27,190 30,040 32,856 36,163 40,062 44,372 49,796 56,321 64,702 75,145 88,064 98,435 108,332 121,098 134,959 147,827 161,577

United States

Japan

India

59,796 61,740 65,393 68,787 73,087 76,640 81,002 86,085 90,891 96,657 102,897 106,253 109,802 115,122 122,770 130,954 138,579 144,803 147,203 144,179 149,583 155,338 162,446 168,000

23,780 25,387 26,179 26,848 27,656 28,782 30,072 31,075 30,783 31,160 32,898 33,771 34,717 35,690 37,534 38,896 40,649 42,643 42,895 40,811 43,227 43,862 45,048 46,244

10,200 10,651 11,491 12,323 13,424 14,742 16,144 17,086 18,339 20,246 21,502 23,055 24,300 26,734 29,643 33,434 37,655 42,443 44,955 49,143 54,841 59,630 63,546 67,744

Data Source World Bank Database, 2014

growth rate was significantly higher than that of the United States. Therefore, the World Bank predicted that in 2014, China was likely to surpass the United States and become the world’s largest economy.2 One dropped pebble can create a thousand ripples. This news immediately sparked a lively discussion among politicians, academics, and business executives; many different views and opinions were heard. However, the Chinese government has been slow to express its views and exhibited little interest in the news. 2 “The IMF’s latest World Economic Outlook report predicts that China will comprise 16.48% of the global GDP at the end of this year, with a scale of USD 17.632 trillion while the United States will comprise 16.28%, with a total of USD 17.416 trillion. By 2019, China’s aggregate economic volume is estimated to be 20% higher than that of the United States” (quoted from People’s Daily on October 11, 2014).

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

3

Within six months after the release of the report, the International Monetary Fund (IMF) and the World Bank announced that when calculated using the PPP method, the economic scale of the United States reached USD 17.4 trillion on October 10, 2014, whereas that of China was USD 17.6 trillion. China has overtaken the United States, ranking first in the world. On October 14, 2014, an editorial in the North American World Journal stated that China would become the world’s largest economy five years earlier than expected. This marks the beginning of the endless disputes over China being the world’s largest economy.

1.2   Long-Standing Debate The ranking of China’s economic scale has become a hot topic. In recent years, many domestic and overseas scholars have discussed China’s ranking in terms of GDP. According to the IMF’s World Economic Outlook, if calculated using the exchange rate method, China’s economic scale ranked 11th in the world in 1990 and overtook Italy and ranked sixth in the world in 2000. In 2003, it surpassed France and ranked fifth in the world, overtook the UK in 2006, ranking fourth in the world, and surpassed Germany and ranked third in the world in 2007. China’s GDP reached USD 5.87 trillion in 2010, surpassing Japan’s USD 5.46 trillion for the first time and becoming the world’s second largest economy after the United States (see Appendix A, Table A.1). When exactly did China overtake Japan and become the world’s second largest economy? Scholars and international research institutions have adopted different methods and reached different conclusions. If the exchange rate method is adopted for calculation, China’s GDP was about USD 5.6114 trillion in 2007, exceeding Japan’s GDP of USD 4.3789 trillion. If the PPP method is adopted for calculation, China’s GDP was about USD 3.3392 trillion in 2001, surpassing Japan’s GDP of USD 3.2933 trillion. According to the data obtained after adjusting the proportion of the service industry, China’s GDP was USD 3.1813 trillion in 1999, already surpassing Japan’s GDP of USD 3.0591 trillion. In any case, it is commonly agreed that China’s economy has surpassed Japan’s economy. Since then, the focus of the discussion has

4  X. SONG

become this: When will China overtake the United States to be the world’s largest economy?3 Professor Yao Yang from China Economic Research Center of Peking University pointed out that a series of assumptions should be made when estimating when China’s economy can catch up with that of America.4 If China’s economy maintains a growth rate of 8.0% while America’s economy maintains a growth rate of 3.0%, China’s inflation rate remains at 3.6% while America’s inflation rate remains at 2.0%, and the appreciation of renminbi (RMB) against USD is 3.0% every year, then the two countries’ GDPs would both be USD 24 trillion in 2021. If China’s economic growth rate is maintained between 9.0 and 10.0% for the first five years and 6.0 to 7.0% in the next five years, then the calculation results would almost be the same, and China will catch up with the United States in 2021.5 In the World Economic Outlook released by IMF in April 2011, the IMF predicted that if PPP is adopted for measurement, China’s economy would overtake the United States in 2016 and become the world’s largest economy. By then, China’s GDP will reach USD 18.7 trillion, whereas America’s will be USD 18.3 trillion. Professor Robert Feenstra from the University of California adopted the income approach to calculate GDP.6 He believed that the World Bank adopted the price index of towns and surrounding area to calculate China’s actual GDP and overestimated the average price level. Therefore, the World Bank perhaps underestimated China’s real GDP by 50.0%. A correction of this indicator would mean that China’s actual GDP will exceed America’s within a shorter period, that is, in 2013 or 2012, rather than 2016 as predicted by the IMF (see Table 1.2). The debate was fruitless due to the dramatically different views. After the World Bank, the IMF, and other authoritative agencies presented their perspectives in October 2014, people began to carefully examine the GDP data to explore the truth. If we look at history, we can find that the United States overtook the UK and became the world’s biggest economy in 1872. The United States has maintained this leading position for 142 years. In the 1950s, 3 See

Chapter 12 of this book. will China’s Economy Overtake America? China Daily, 2012. 5 See Yao Yang, When Will China’s Economy Overtake America’s? China Daily, 2012. 6 See Robert Feenstra (2012), “How Big Is China?” China Economic Quarterly 11, vol. 2 (2012): 367–382. 4 When

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

5

Table 1.2  Comparison of Sino-US economic scale calculated by American Professor Stephens (Unit: Trillion USD) Country

2005

2008

2011

2012

2013

United States China

123,640 68,630

127,160 89,160

130,780 115,830

132,010 126,390

133,250 137,900

Data Source Robert Feenstra, “How Big Is China?” China Economic Quarterly 11, vol. 2 (2012), p. 367

when the Chinese declared the goal to “surpass the United Kingdom and catch up with United States,” many people viewed this slogan as a mere romantic fantasy. The fact that China has overtaken the United States as the world’s largest economy surprised many countries and many Americans found it hard to accept, even the Chinese may find themselves unprepared for this development. According to a report of the Asian Development Bank (ADB), the Chinese government’s attitude can be summarized as follows: “the National Bureau of Statistics (NBS) of [the People’s Republic of] China expresses reservations over some aspects of the methodology employed and does not agree to publish the headline results for the People’s Republic of China (PRC). The results for the PRC are estimated by the 2011 ICP Regional Office in the Asian Development Bank and the 2011 ICP Global Office in the World Bank. The NBS of [the People’s Republic of] China does not endorse these results as official statistics.”7 Some people deemed the data calculated by World Bank using the PPP method as “unreliable” and should be ignored, some questioned the accuracy of the data provided by the World Bank, some thought that the World Bank wanted to flatter China, and some considered such arguments as having ulterior motives and building a basis for the “China threat theory.” Nonetheless, most scholars believed that finding fault with China would not be worthwhile for the World Bank. After all, it is an international organization. In addition, the raw data used by the World Bank were provided by the official statistical agencies of China, and the Chinese government sent people to participate in the ICP. There must be inherent reasons behind the conclusions of the World Bank and

7 See the Asian Development Bank, “2011 International Comparison Program in Asia and the Pacific. Purchasing Power Parities and Real Expenditures: A Summary Report”, p. 12.

6  X. SONG

the IMF. Whether people accept or reject the views of the World Bank, they must face the challenges and conduct serious, in-depth analysis and research. It is time to address the issue directly.

1.3   Difficulties in Measuring Economic Scale Why is China’s economic scale highly disputed? It is commonly known that measuring length requires a ruler, and the scale of all rulers must be the same. At present, the Chinese use the metric system (e.g., meter), while the Americans use the imperial system (e.g., foot). Although the figures obtained are different, fixed rules are used to convert between the units to avoid misunderstanding. In fact, the Chinese used to have chi as the basic unit of length and later changed to meter for the ease of communicating with other countries. One meter equals three chi. However, in ancient Chinese literature, Guan Gong8 was supposedly  nine chi tall. If converted into meters, he would have been three meters tall. Obviously, this is not possible. The reason behind this confusion is that the length of one chi during the Han Dynasty was much smaller than the current chi. Although people have been using “chi” as a unit of measurement, its connotation has changed considerably with the passage of time, resulting in difficulties in understanding ancient literature. Thus, only through standardized measurement can we prevent misunderstandings in the communication process. Measuring an economy’s scale is more difficult than measuring an object’s size or weight because it involves not only the quantity and quality of the products, but also the conversion between the purchasing power of the domestic and foreign currency. Measuring industrial and agricultural products is very complex, and the contents of the service sector are even more complicated. Hence, improving the accuracy of the domestic measurement is already difficult, not to mention the comparison among countries. Every country in the world has its own statistical system and calculates the GDP of all sectors and regions every year. The most primitive domestic product data are calculated using the local currency. The fundamental purpose of each country’s GDP calculation is to collect taxes. Tax revenue is indispensable in maintaining the normal running of the 8 Guan Gong, formerly known as Guan Yu, was a general of the Three Kingdoms period in ancient China and was called “Guan Gong” by later generations.

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

7

machinery of government (e.g., military, courts, police, fire departments, etc.). Without basic statistics, how can each country reasonably collect taxes to meet the financial needs of government administration, social welfare, infrastructure, military defense, and other aspects? Finance and taxation should be allocated according to the product and real incomes of different sectors, residents, and regions. Most raw statistical data come from the taxation process. People often use national statistical data to analyze the economic growth rate, structual change, income gap, infrastructure investment, etc.

1.4  Two Sets of GDP Statistical Data With the development of international trade and international finance, people should conduct horizontal comparisons among countries. The easiest approach is to multiply the GDP obtained by each country using an exchange rate, convert all GDPs into the same currency, and then perform comparisons (see Table 1.3). This methodology is commonly known as the exchange rate method. Taking CNY as the unit, China’s GDP and GDP per capita are shown in the second and third columns of Table 1.3, respectively. Although these data may have some errors, they are at the same time original and irreplaceable. The fourth column of the table presents the average annual exchange rate for each year. For example, the USD to RMB exchange rate was 1:6.19 in 2013. Based on exchange rate, the GDP that takes CNY as a unit can be converted into USD. The fifth and sixth columns of the table present China’s GDP and GDP per capita when USD is taken as the calculation unit, respectively. Careful readers may find these figures slightly different from the figures published by several official agencies. In fact, the data released by various international statistical agencies have huge differences. Although the exchange rate is used as the conversion factor between RMB and USD in all cases, results are different due to the adoption of the mid-year exchange rate, year-end exchange rate, average annual exchange rate, or the average of the floating exchange rates in recent years. The fluctuation range of the exchange rate is limited, but the GDP calculated using different kinds of exchange rate may differ significantly from one another. Therefore, choosing the conversion factor is important when performing international comparisons.

8  X. SONG Table 1.3  China’s GDP calculated according to the exchange rate method Year

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

GDP (100 million CNY)

9016 10,275 12,059 15,043 16,992 18,668 21,781 26,923 35,334 48,198 60,794 71,177 78,973 84,402 89,677 99,215 109,655 120,333 135,823 159,878 184,937 216,314 265,810 314,045 340,903 401,513 473,104 518,942 568,845

GDP per capita (CNY)

858 963 1112 1366 1519 1644 1893 2311 2998 4044 5046 5846 6420 6796 7159 7858 8622 9398 10,542 12,336 14,185 16,500 20,169 23,708 25,608 30,015 35,198 38,420 41,908

Average annual exchange rate: CNY/1USD 2.94 3.45 3.72 3.76 4.78 5.32 5.51 5.76 8.62 8.35 8.31 8.29 8.28 8.28 8.28 8.28 8.28 8.28 8.28 8.28 8.19 7.97 7.60 6.95 6.83 6.77 6.46 6.31 6.19

GDP (100 million USD)

3070 2976 3240 3996 3552 3507 3950 4673 4100 5772 7312 8586 9539 10,196 10,833 11,987 13,248 14,538 16,410 19,316 22,576 27,135 34,957 45,218 49,905 59,312 73,250 82,209 91,850

GDP per capita (USD)

292 279 299 363 318 309 343 401 348 484 607 705 775 821 865 949 1042 1135 1274 1490 1732 2070 2652 3414 3749 4434 5450 6086 6767

Data Source China statistical abstract

Because the exchange rate method has many shortcomings, since the 1980s, institutions such as the World Bank and the IMF have published two kinds of GDP statistics simultaneously, namely, GDP obtained with the exchange rate method and GDP obtained with the PPP method.9 The United Nations, the World Bank, and other international organizations 9 Please refer to Chapters 3 and 4 for the analysis on the shortcomings of the exchange rate and the purchasing power parity methods.

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

9

have repeatedly expressed that GDP data obtained with the PPP method is still somewhat problematic, and should only be used as reference for research and observation. In recent years, the World Bank and the ADB have organized a strong team and allocated considerable manpower and material resources to improve the PPP method. After years of efforts, the PPP method has gained considerable progress. Data recently released by the World Bank were estimated with the PPP method. A comparison between the two kinds of GDP data—one calculated with the PPP method and one with the exchange rate method—reveals considerable differences between the two algorithms (as shown in Table 1.4). For example, in 1990, China’ GDP was USD 1.1427 trillion Table 1.4  Gaps of GDP between the two methods Year

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Purchasing power parity method (100 million USD) 11,427 12,260 13,691 16,012 18,637 21,514 24,301 27,190 30,040 32,856 36,163 40,062 44,372 49,796 56,321 64,702 75,145 88,064 98,435 108,332 121,098 134,959 147,827 161,577

Exchange rate method (100 million USD) 3507 3950 4673 4100 5772 7312 8586 9539 10,196 10,833 11,987 13,248 14,538 16,410 19,316 22,576 27,135 34,957 45,218 49,905 59,312 73,250 82,209 91,850

Ratio of GDPs derived from the two methods 3.26 3.10 2.93 3.91 3.23 2.94 2.83 2.85 2.95 3.03 3.02 3.02 3.05 3.03 2.92 2.87 2.77 2.52 2.18 2.17 2.04 1.84 1.80 1.76

Data Source Data for PPP method are from the World Bank database, data for exchange rate method are from Table 1.3

10  X. SONG

when calculated with the PPP method, but was only USD 350.7 billion when calculated with the exchange rate method. The former is 3.26 times as big as the latter. That is to say, the exchange rate method may underestimate China’s GDP. And yet, in other words, the PPP method is likely to overestimate China’s GDP. However, with the development of economies, the gap between the two statistical methods gradually narrowed from 3.26 times in 1990 to 1.76 times in 2013. Globally speaking, this problem is not unique to China but is a common problem faced by almost all developing countries. The data obtained using the two statistical methods have huge differences. The gap appears to be larger in economically underdeveloped countries. For developed countries, the gap in the data obtained through these two statistical methods is very small and no significant difference can be observed. However, for developing countries, the conclusions derived from these two methods may differ more significantly, which could easily lead to misunderstandings.

1.5   Different Statistical Conclusions Can Be Reached Through Different Observation Perspectives Which kind of statistical method can produce reliable statistical data: the exchange rate method or the PPP method? Obviously, if one of the two was more comprehensive and reasonable, the other method would have been abandoned long time ago. The reason why these two statistical methods coexist is that both methods are imperfect. That is to say, they each have advantages and disadvantages. Using any of the two methods to measure and observe GDP is reasonable. Different statistical conclusions can be reached with different observation perspectives. The measurement problem can be very complex when discussing the economic size of a country. The world is complicated and ever-changing. Quantifying economic activities is difficult. Some goods and production factors cannot stride over national borders and form a unified market, and some elements cannot enter trade. Without adequate market mechanisms, no reasonable price will exist. Because economic statistics cannot consider everything, inevitable omissions will exist. Some problems are difficult or impossible to solve, which is why everlasting disputes over the scale of China’s GDP exist, but no conclusion has been reached.

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

11

More importantly, what is the purpose of comparing international economic scales? While some researchers wish to study, observe, and compare a country’s overall national capacity, some wish to investigate the income, welfare, and the wealth and poverty among residents, and others focus on international trade and international finance. If people want to study a country’s comprehensive national capacity, they will naturally focus on material products. Researchers seeking to examine the wealth and poverty will naturally pay more attention to GDP per capita, income, and consumption level. Those who wish to study international trade and finance need to discuss the exchange rate in the transaction process. Different research purposes determine different observation angles. For example, sprinting and swimming competitions attach great importance to speed; weightlifting competitions attach great importance to weight; and football and basketball games attach great importance to teamwork. Different sports require different skillsets and have different competition rules, and judgment standards. Therefore, when discussing the economic size of a country, a single statistical method should not be used, and no single statistical method exists. Hence, a thorough discussion of all relevant GDP statistical methods and an in-depth study of the history of China’s GDP statistics and previous adjustments of GDP data are necessary. People should know the Chinese characteristics in GDP statistics, i.e., China’s tax system differs significantly from that of the western countries, which causes massive loss of data in the service sector. The exchange rate method and the PPP method are two statistical methods that measure a country’s economic size from different perspectives. The two methods draw different conclusions. Arguing about when China will overtake the United States to become the world’s largest economy is unnecessary. If people shift their angle of observation, they may find that China lags far behind the United States, China cannot surpass the United States within the near future, or China will overtake the United States and become the world’s largest economy much earlier than the time announced by the World Bank and IMF. Only by taking calm, rigorous, objective, and fair observations can people know the actual size of China’s GDP, constantly adjust relationships with other countries in the process of rapid economic development, and promote peace and development.

12  X. SONG

1.6  The World Bank’s Two Sets of Rankings According to the data released by the World Bank, at least two sets of rankings show the economic scale of various countries in the world, one using the PPP method and one using the exchange rate method. The recent ranking list published by the World Bank used the PPP method in accordance with the International Comparison Program (ICP). The top 30 largest economies by GDP measured in PPP terms are shown in Appendix A, Table A.1. For comparison, the top 30 largest economies by GDP measured in exchange rate terms are shown in Appendix A, Table A.2. The results using the PPP method (see Table 1.5) have shown that, in 1990, the United States ranked No. 1 with a GDP of USD 5.9796 trillion; Japan ranked No. 2 with a GDP of USD 2.3780 trillion; Germany ranked No. 3 with a GDP of USD 1.4721 trillion; Russia (then the Soviet Union) ranked No. 4 with a GDP of USD 1.1894 trillion; China ranked No. 5 with a GDP of USD 1.1427 trillion; and India ranked No. 6 with a GDP of USD 1.0200 trillion. Taking the GDP of the United States as reference, Japan’s economic size was 39.8% of that of the United States, Germany’s economic size was 24.6% of that of the United States, Russia’s economic size was 19.8% of that of the United States, China’s economic size was 19.1% of that of the United States, and India’s economic size was 17.1% of that of the United States. China’s economic size was similar to that of India. The results using the exchange rate method (see Table 1.6) have shown that, in 1990, the United States ranked No. 1 with a GDP of USD 5.9796 trillion; Japan ranked No. 2 with a GDP of USD 3.1037 trillion; Germany ranked No. 3 with a GDP of USD 1.2402 trillion; China ranked No. 11 with a GDP of USD 0.3569 trillion; and India ranked No. 12 with a GDP of USD 0.3266 trillion. Taking the United States’ GDP as reference, Japan’s economic size was 51.9% of that of the United States, while China’s economic size was 6% of that of the United States. In 2013, the GDP rankings measured in PPP terms underwent significant changes. The GDP of the United States was USD 16.8 trillion, whereas China’s GDP was USD 16.1577 trillion. The two countries almost had the same GDP size. As a result, the World Bank concluded that China would overtake the United States in 2014 and become the

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

13

Table 1.5  Horizontal comparison of GDP measured in purchasing power ­parity terms (Unit: 100 million USD) Year

China

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

11,427 12,260 13,691 16,012 18,637 21,514 24,301 27,190 30,040 32,856 36,163 40,062 44,372 49,796 56,321 64,702 75,145 88,064 98,435 108,332 121,098 134,959 147,827 161,577

United States Japan

59,796 61,740 65,393 68,787 73,087 76,640 81,002 86,085 90,891 96,657 102,897 106,253 109,802 115,122 122,770 130,954 138,579 144,803 147,203 144,179 149,583 155,338 162,446 168,000

23,780 25,387 26,179 26,848 27,656 28,782 30,072 31,075 30,783 31,160 32,898 33,771 34,717 35,690 37,534 38,896 40,649 42,643 42,895 40,811 43,227 43,862 45,048 46,244

US GDP as the benchmark (%)

Chinese GDP/US GDP (%)

Japanese GDP/US GDP (%)

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

19.1 19.9 20.9 23.3 25.5 28.1 30.0 31.6 33.1 34.0 35.1 37.7 40.4 43.3 45.9 49.4 54.2 60.8 66.9 75.1 81.0 86.9 91.0 96.2

39.8 41.1 40.0 39.0 37.8 37.6 37.1 36.1 33.9 32.2 32.0 31.8 31.6 31.0 30.6 29.7 29.3 29.4 29.1 28.3 28.9 28.2 27.7 27.5

Data Source World Bank Database, 2014

world’s largest economy because China’s GDP growth rate was significantly higher than that of the United States. If calculation was performed using the exchange rate method, in 2013, China’s GDP was USD 9.2403 trillion, which accounted for only 55.0% of the US GDP (USD 16.8 trillion). China’s economic size was only a little more than half of that of the United States, lagging far behind the world’s largest economy. India’s ranking in terms of economic scale is also noteworthy because it continues to rise. In 1990, India’s GDP was USD 1.02 trillion,

14  X. SONG Table 1.6  Horizontal comparison of GDP measured in purchasing power ­parity terms (Unit: 100 million USD) Year

China

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

3569 3795 4227 4405 5592 7280 8561 9527 10,195 10,833 11,985 13,248 14,538 16,410 19,316 22,569 27,130 34,941 45,218 49,902 59,305 73,219 82,295 92,403

United States Japan

59,796 61,740 65,393 68,787 73,087 76,640 81,002 86,085 90,891 96,657 102,897 106,253 109,802 115,122 122,770 130,954 138,579 144,803 147,203 144,179 149,583 155,338 162,446 168,000

31,037 35,368 38,528 44,150 48,503 53,339 47,062 43,243 39,146 44,326 47,312 41,599 39,808 43,029 46,558 45,719 43,567 43,563 48,492 50,351 54,954 59,056 59,378 49,015

US GDP as the benchmark (%) 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Chinese GDP/US GDP (%) 6.0 6.1 6.5 6.4 7.7 9.5 10.6 11.1 11.2 11.2 11.6 12.5 13.2 14.3 15.7 17.2 19.6 24.1 30.7 34.6 39.6 47.1 50.7 55.0

Japanese GDP/US GDP (%) 51.9 57.3 58.9 64.2 66.4 69.6 58.1 50.2 43.1 45.9 46.0 39.2 36.3 37.4 37.9 34.9 31.4 30.1 32.9 34.9 36.7 38.0 36.6 29.2

Data Source World Bank Database, 2014

accounting for 17.1% of the United States’ and 42.9% of Japan’s GDPs. In 2007, India’s GDP was USD 4.24 trillion, whereas Japan’s GDP was USD 4.26 trillion. The sizes of the two economies were almost the same. In 2008, India overtook Japan and became the world’s third largest economy. Japan was relegated to No. 4 from No. 2. In 2013, India’s GDP was USD 6.77 trillion, whereas Japan’s was USD 4.62 trillion. Japan’s GDP accounted for only 68.2% of India’s GDP. According to the PPP method, the GDP ranking of many developing countries with large populations underwent tremendous changes

1  ONE DROPPED PEBBLE CREATES A THOUSAND RIPPLES 

15

over the period 1990–2013. For example, Indonesia’s rank moved up from No. 14 to No. 9 and surpassed that of the UK, Italy, and Canada. Brazil’s rank rose from No. 9 to No. 7, whereas Germany’s rank dropped from No. 3 to No. 5. The following questions should be answered when discussing China’s economic scale: 1.  Although Chinese authorities have some reservations about the data measured in PPP terms, the World Bank and ADB continue to release their calculation results and the detailed calculation formulas and procedures according to the PPP method. Hence, what is the PPP method? Compared with the commonly used exchange rate method, what are the advantages and disadvantages of the PPP method? 2.  What can be done to improve and develop the PPP method adopted by the World Bank and other international organizations? 3. How should we view the international GDP ranking? We will discuss all these questions in the following chapters.

CHAPTER 2

Re-examining Some Problems in the Methodology of International Economic Comparison

The results of the eighth International Comparison Project (ICP), which was conducted by the World Bank together with other countries and organizations, were published in 2014, and they reignited the debates on various aspects of the project. Economists have divided perspectives on this project. Most experts engaged in international comparison approve of the project and hold its basic idea and comparative framework in high regard, and they believe that the technical details of its comparison method should be improved further. Some economists, however, are generally suspicious of the methodology and results of the ICP. Although we highly respect the creative contributions of ICP experts, we acknowledge that the ICP has a problematic methodology that requires further investigation. We also insist that the exchange rate method has several advantages over the ICP method. This ­chapter focuses on four aspects, namely, the premise setting of comparison homogeneity, the rethinking of the comparison between the ICP and the exchange rate methods, the comparison of spatial structure and its measures, and the need for further study on the subject of international economic comparison.

2.1   Premise Setting of Comparison Homogeneity Trading signifies the unity of opposites between commodities and currencies. This practice is manifested as the relative relationship between the purchasing power of currency and the price of commodity in the © The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_2

17

18  X. SONG

same unit. A low price implies strong currency, many purchasable ­commodities, and high purchasing power, whereas a high price implies weak currency, few purchasable commodities, and low purchasing power. The price is inversely proportional to the purchasing power of currency. The total value of a commodity includes three factors, namely, quantity, quality, and price. After controlling for quantity and quality, the price can be used to compare purchasing power. If commodities are measured with unit price, the quantity factor is eliminated. If these commodities are confirmed to be similar, the quality factor is eliminated. Therefore, the unit price ratios of commodities can reflect the purchasing power ratio of currencies. In this case, the homogeneity of commodities is a very important assumption for comparing the purchasing power of different currencies. Commodity homogeneity has microscopic and macroscopic levels. The former pertains to the homogeneity of expenditure, and the latter pertains to the homogeneity of the economies in comparison. 2.1.1   Homogeneity of Expenditure The price ratio of homogeneous expenditure reflects the purchasing power of the currency. If the expenditure is heterogeneous, then the quality factor is involved and the price ratio cannot truly reflect the purchasing power of the currencies. However, homogeneity is difficult to achieve in economic reality. The Economist has invented the “Big Mac Index” to reflect the purchasing power parity (PPP) with a single product, which was highly appraised by many economists. On the surface, the Big Mac, a hamburger sold at McDonald’s restaurants can be found in any McDonald’s in the world, and its production and sale follow a standardized procedure. Therefore, the Big Mac has high comparability and is sufficient to reflect the PPP among different currencies. However, the Big Mac Index is not free of error. This is because that deducing the PPP using a single product is absurd, not to mention that Big Mac in different countries is not exactly the same. For instance, hamburger is just fast food in Europe and North America, however, in less-developed economies, it is a symbol of culture and modern life, a fashionable lifestyle, an economic behavior that claims open-mindedness and passion for life, and even a preferable choice of reward from parents. In other words, a Big Mac carries much richer cultural meaning in less-developed economies than in developed economies.

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

19

The homogeneity of commodities can be determined from different perspectives, such as output, input, process, and integration. What makes one perspective better than another? Any choice should be supported by a sufficient number of arguments. If the inputs are different, the outputs will be different even measured with a unified standard. For example, given that the inputs of a Big Mac hamburger (i.e., meat, flour, potato, oil, tomato, wrapping paper, and labor) are different, one cannot say that all Big Mac hamburgers are entirely comparable. No man ever steps in the same river twice, and no two things in the world are entirely the same. The so-called homogeneity can only be relative. Determining the homogeneity of products or expenditure has remained a difficult econometric problem for humans throughout history. ICP emphasizes the comparison of volumes of final goods and services between economies and follows the so-called real comparison principle, which is an extension of the homogeneity principle or a stratified version of it. This principle serves as a working standard that focuses on feasibility. Although the production method can also be adopted when making international economic comparisons, the expenditure approach is usually adopted in these cases because of its ability to confirm homogeneity. Given the relative homogeneity or homogeneous approximation, choosing the comparison method and applying the comparison conclusion should be done in a prudent manne. We should not forget or ignore the fact that in the real world, a potato is not merely a potato. Even products with the same physical form cannot be classified as homogeneous. Different sale locations and conditions bring different economic benefits and quality of life. In other words, although two products may seem similar, their internal qualities may greatly vary. Such variances are mainly attributed to the differences between developed and developing countries. For example, developed countries provide after-sales services such as return and exchange when selling commodities. Clearly, if these service agreements are absent, the product price will sharply decrease. By contrast, commodities in developing economies are cheap not because of the strong purchasing power of their currencies but because the commodity includes fewer services. Similarly, the high prices of commodities in developed economies do not imply the weak purchasing power of their currencies but the fact that such commodities include many services. Ignoring the difference in the qualities of these products overestimates the purchasing power of the

20  X. SONG

currencies of developing countries and underestimates that of developed countries. Economic development trends should also be considered. Personal preference has an important influence on price. Customization lowers the homogeneity of economic projects and increases the difference in the accompanying services. Customized production creates the opposite force against economic homogeneity. The existence of such a factor can guarantee the vitality of the economy but reduces the homogeneity of the output or input. Homogenous and heterogeneous orientations both exist in economic reality. Therefore, what is the influence of this trend on international economic comparison? The customized development of products can at least prohibit the promotion of the ICP method. 2.1.2   Homogeneity of Economies in Comparison The homogeneity of comparable items in operation should be determined. However, are all expenditures comparable? ICP requires the commodities in a specific market to be of high variety so there are possible comparable items. This requirement tells us that there are incomparable commodities in different economies, and homogeneity in this case is a macro-level problem. The differences among the commodities of various countries should not be neglected. The markets of some countries have special commodities that are considered rare in other countries. In other words, the purchasing power of Country A’s currency is attributed to its special commodity a and common commodity c, while the purchasing power of Country B’s currency is attributed to its special commodity b and common commodity c. Given such market differences, if the comprehensiveness requirement must be satisfied to calculate PPP, one must compare the price parities between special commodities a and b. However, such parities are not observed in economic reality. That is, the total price parity of a currency’s purchasing power contains blank subsets, and thus ICP has some incomparable objects and compares some non-existent expenditures. Do these findings suggest that ICP goes beyond the measure boundary? To address this problem, the ICP has adopted the substitution method, in which the comparable commodities are taken and a representative commodity is substituted for incomparable ones. This method focuses on comparable expenditures, directly calculates the price parity,

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

21

and assumes that the price ratio of non-representative commodities is equal to that of the representative commodity. Given the subdivision of spending categories, the products in one category only show a slight difference from those in another category. The non-representative commodities do not largely deviate from the representative commodities. If the non-representative commodities have a price ratio, this ratio will also not largely deviate from that of the representative commodity. Therefore, the assumption of this substitution method seems reliable. Substituting incomparable products with representative ones assumes economic isomorphism. Economies with similar levels of development are more suitable to perform ICP, which is a manifestation of the homogeneous requirement. In this case, what should be done for incomparable products? Merely stating that the comparable items are “considerably overlapping” is not enough; one should also determine how many items are overlapping, why having more overlapping items is better, how many items should have overlap to satisfy necessary conditions, and whether different degrees of overlap among countries indicate differences in the reliability of their ICP results. The incomparable part of the expenditure is a direct embodiment of different economic structures. Therefore, the homogeneity and heterogeneity of the basket of goods in different countries should be calculated according to their number of product varieties or amount of expenditures to reflect the reliability of the ICP results. In the ICP processing, the differences between non-representative and representative commodities are not eliminated but are merely decomposed and confined in a sub-category. When we sum up the price ratios of the representative commodities (equalization), the differences are also aggregated. What is the effect of using representative commodities as substitutes in different economic structures? This question needs to be further examined. During its early stages, the ICP was only performed in countries with developed market economies to obtain accurate calculations. However, in its later stages, the ICP was performed in countries all over the world and the effect of different economic structures deepens. Can the calculation model derived from a competitive market economy effectively reflect the operation of an incomplete market economy? Is it too dangerous to use one model to explain different economic realities? These problems should be considered when evaluating the accuracy of the ICP results.

22  X. SONG

The eighth round of the ICP had the largest number of participating economies and also compared the most heterogeneous economies in the world. The ICP has a fundamental dilemma: On the one hand, only by including more economies in the comparison can the results truly reflect the global economic scale; on the other hand, the countries that participated in the ICP during its later stages demonstrated strong heterogeneity, which make it more difficult to achieve accuracy.

2.2  Rethinking the ICP and Exchange Rate Methods Why was the ICP method developed when the exchange rate method could also be used for international comparison? The ICP method is intended to measure the purchasing power of different currencies more accurately and to measure the real economic scale at the economic, regional, and global levels. Therefore, comparing the ICP and exchange rate methods becomes inevitable. Most scholars working on international economic comparison prefer the ICP method and have criticized the exchange rate method, and many have incorporated the ICP method into their work either explicitly or implicitly. However, in the latest IPC methodology manual, introductions to these two methods have changed. That is, the manual no longer rejects the exchange rate method. These changes reflect a scientific attitude and suggest that we should be more open-minded. Are the results of the ICP method better than those of the exchange rate method? In other words, to what extent does the ICP method reflect economic realities? Are the results from the exchange rate method entirely useless? These problems warrant further discussion. 2.2.1   Orientation of the Validity of the International Comparison Results The exchange rate is the basic scale for international trade in reality. Anyone who exchanges currencies when traveling abroad cannot adopt the comparison results of the ICP as his/her standards but must refer to the market exchange rate. After the latest ICP results were published, some people commented that China should not purchase missiles, ships, mobile phones, and German cars with price ratios marked by the PPP

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

23

but should pay for them according to the exchange rate. Therefore, the exchange rate is extremely important when comparing the strengths of different countries. China is not asking for a special treatment when, in fact, all countries prefer such treatment. Clearly, the market exchange rate is a better measure for the purchasing power of the people and enterprises in the international market. Therefore, the validity of the exchange rate method has a microeconomic foundation. This method is based on the actual practices of daily international trade and also individual trading items. The results of the exchange rate method may be more valid than those of the ICP method, especially with regard to the currency’s purchasing power. How does the validity of the ICP method results manifest itself? The author believes that it should be primarily on the macro—and average level. That is, the results of the ICP method are not superior to those of the exchange rate method in all aspects. Many people argue that the exchange rate does not reflect the relative purchasing power of currencies of different countries in their own markets, as it only reflects the proportional relationship between the prices of goods and services in international trade. In the direct sense, this view is correct; but in the indirect sense, the exchange rate reflects some aspects of the purchasing power. The notion that the exchange rate method only considers the price ratio of traded items is superficial. As we all know, the domestic market is connected with the international market, and their prices are not isolated. Instead, the prices of different commodities can influence one another and are mutually dependent. The formation of the prices of trading and non-trading commodities is an interconnected process. Therefore, why does currency speculation exist? Isn’t it a result of the differences in the purchasing power of currencies? The direction and the degree of the speculation depend on that of the purchasing power disparity. The exchange rate method, which reflects the price ratios of trading commodities, can also indirectly reflect the price ratios of non-trading commodities to a certain extent. Therefore, given its indirect influence, stating that the exchange rate method is useless is unfair. Indeed, as a comprehensive method, the exchange rate method cannot distinguish (decompose) the price ratio relationship of a single item, but this does not indicate that this method does not consider such a relationship.

24  X. SONG

2.2.2   How to Understand the So-Called Political Distortion of Exchange Rates The results of the exchange rate method are often questioned because the exchange rate is easily affected by non-economic factors. That is, governments may intervene with the exchange rate, and such political distortion leads to signal distortion. The following points should be considered: The first point pertains to the allelism of exchange rate and price distortions. The exchange rate is a comprehensive price of the international currency market and is among the many types of prices in such a market. The force of the government on the economy is not one dimensional. If the exchange rate is distorted, the price is also distorted. A perfect market with full competition does not exist, and the price can never perfectly reflect the supply and demand relationship. The second point pertains to distorted multi-subjects. If price distortion exists, both politica—and monopoly-driven (i.e., intervention of monopoly enterprises in the price) distortions will be observed in actual economic situations. Both the government and the monopoly enterprises intervene in the market price, and even non-government organizations (NGOs) and foreign entities (i.e., foreign residents, enterprises, governments, and NGOs) can greatly influence the price decision. The System of National Accounts (SNA) recognizes five economic subjects, namely, residents, enterprises, government, NGOs, and foreign entities, which are all formed in the economic process. Price is formed in the gaming of these five economic subjects. In modern international economic comparison, failure to acknowledge that other economic subjects except for residents and enterprises also exist and to consider the influence from the other economic subjects on price decisions as external intervention is inconsistent with the logical framework of the SNA and the basic idea of its structural design. The third point pertains to the degree and level of distortion. If government distortion exists, the exchange rate distortion will be reflected in a concentrated form. The exchange rate method accepts a generalized distortion. However, price distortion is reflected in a dispersed form. The ICP price collection accepts the distortions in the prices of real expenditures in the process of collecting data on market prices, and then aggregates all distortions in the analysis. In this case, the ICP price collection method only differs from the exchange rate method in terms

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

25

of its degree and level in receiving such distortions. Therefore, the question becomes the following: On what level are these distortions directly or indirectly received? Some people propose that the distortions in price information can offset one another. If that is true, the distribution patterns of different distortions should be examined. The fourth point pertains to the “distortion of the distortions.” Government intervene in price decisions for both political and economic reasons. The economic history of many countries reveals that the so-called night watchman theory is simply a myth, and the “List’s ladder” always exists under the “loft” of economic development. The degree of participation of the government in the economy may differ under various development stages. For a country with a weak economy, government intervention in prices is a necessary means to cope with the intervention from more powerful economies and (transnational) monopoly enterprises. Therefore, such an intervention may play the role of positive adjustment rather than distortion. In other words, it is a counterforce to distortion, and namely, the “distortion of distortions.” For example, does minimum wage contribute to price distortions? Are a government’s regulations on the wage negotiation between enterprises and trade unions positive adjustment to price distortions? Is the implementation of anti-monopoly laws considered a form of anti-price distortion? Can all influences of supra-governments, such as the World Bank, the International Monetary Fund, and other international organizations, on national economies boil down to price distortion? As the ICP itself is a government behavior, can this project be regarded as a distortion of PPP? For these four reasons, the so-called government distortion cannot be used as a valid reason for rejecting the exchange rate method. In this aspect, the ICP method is not necessarily superior to the exchange rate method. If the government’s intervention in the exchange rate is unacceptable, its intervention in the price should not be accepted either. 2.2.3   Limitation of the Effective Space of the Precise Method In terms of precision and complexity of design, the ICP method is undoubtedly superior to the exchange rate method. However, this advantage is not necessarily maintained. Methods with greater technical complexity are not necessarily better. Similarly, the ICP method is not

26  X. SONG

necessarily superior to the exchange rate method. We cannot equate method advantage with result advantage. People often favor complexity over simplicity and are willing to believe in and adopt the method with higher level of technical complexity. When evaluating a measure, they assess the reliability of the results based on the superiority of the method. The general logic is that a more precise method leads to more accurate results. In fact, the precision of a method is a necessary yet insufficient condition for obtaining accurate results. If the premise of the necessity cannot be satisfied, the precise method may give a wrong result. When comparing different methods and their results, people often forget the hypothetical assumptions implied by an improved or innovative method or do not realize the trap of technical innovation. In this case, the precise method is worse than the rough method because the former can easily lead people to believe the wrong results. That is, the inclination of people for technolatry intensifies the negative effects of the method improvement trap. Methods designed with advanced technology always has more hypothetical premises than simpler methods to ensure accurate results. Generally, the more advanced a technology is, the more assumptions are made. However, in measurement reality, data collection does not necessarily meet hypothetical premises. To fulfill or complete the measure, the assumptions must be relaxed. From the perspective of space, a more advanced technology requires more hypothetical premises but has a narrower effective application space. This mismatch between the effective space of the method with advanced technology (small space) and the real space (large space) causes the method to malfunction outside of its effective space. However, people always ignore the difference between these spaces and only apply the method that is effective in a certain space to its malfunction space. They blindly accept the results, thus fall into the trap of method improvement or innovation. In terms of international economic comparison, the ICP needs to compare gross domestic product (GDP) expenditures item by item and ensure these items’ homogeneity, which is very difficult to achieve. In addition, adopting the economic statistics of different countries increases the types and chances of errors. Including more countries can generate more representative results but makes it more difficult to ensure comparability. Comparatively speaking, even if the exchange rate method faces such a problem, it is less obvious than that in the ICP method.

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

27

2.2.4   Comprehensive Comparison Between the ICP and the Exchange Rate Methods Some economists still reject the exchange rate method because of constantly changing exchange rates (when, in fact, price also changes all the time). They seem to ignore the smoothing work of the World Bank ATLAS method and the progress of the exchange rate method. A very important reason for its rejection is that the exchange rate method may generate strange results that cannot be reasonably explained. However, the ICP method may also generate strange results. Neither the generalized nor the deductive descriptions of the ICP method can ensure that the strange results will never appear. Once the measurement model is determined, the calculation process is partially in a black box and the researcher cannot ensure results meet the theoretical expectations. The exchange rate refers to the price ratio of currencies exchanged in the international market. When adopting the exchange rate method, the market indicators are recorded, and the necessary adjustments are applied in the trading records accordingly. Compared with the ICP method, the exchange rate method can only directly compare the international trading commodities. It makes the general comparison, as it belongs to the comprehensive comparison method. The calculation of PPP by the ICP begins from single items (commodity and service) and performs a recursion layer by layer to determine the price ratio relationship among different countries through a set expenditure system. This bottom-up, sub-item comparison follows the structure comparison method. Since the ICP uses indicator in its calculation, the PPP can be seen as the weighted average of the price comparisons of the same expenditures across different economies (countries or regions). Therefore, the ICP method has a wider comparison scope (including all GDP expenditures) than the exchange rate method. Given the macroeconomic significance of the ICP method, PPP is an economic analysis indicator that is artificially calculated and can be classified as an artificially constructed index. The basic settings of the mode of thinking and index calculation show great consistency between the ICP and the exchange rate methods. Fundamentally, both methods take the average of prices in comparison. Given the comprehensive property of exchange rates, the exchange rate method demonstrates the averaging treatment that is inherent in

28  X. SONG

the economic process, whereas the ICP method demonstrates an artificial averaging treatment that is external to the economic process. When measuring the price ratio, the exchange rate method essentially sets the price ratio of non-trading commodities as equal to that of trading commodities, whereas the ICP method assumes that the price ratio of non-representative commodities is equal to that of representative commodities. By comparing their numerical sequences, the calculation results of the two methods are highly correlated, especially for developed economies. Theoretically, the ICP method is more suitable for economies with developed markets. However, if structural data are not considered, only a small difference can be observed in the results between the PPP and the exchange rate methods for developed economies. In this case, the ICP method is unnecessary. In addition, the application of the ICP method actually faces a paradoxical situation. If the exchange rate method is accepted, the existence of the ICP seems meaningless. After all, the ICP method was developed to make up for the flaws in the exchange rate method. However, the ICP method is not completely independent from the exchange rate. In other words, the results of the exchange rate method cannot be completely rejected for the sake of applying the ICP method. A most obvious example is the setting of the price level index (PLI). The ICP methodology manual defines PLI as the comparison between the PPPs of the exchange rate and the ICP methods. The relationship among the PLI, the exchange rate method, and the ICP method indicates that if the PLI and the PPP of the ICP method are approved, the results of the exchange rate method should also be approved. Logically, if the results of the exchange rate method are considered as distorted, such distortion also exists in the results from the ICP method, either its PPP or the PLI. Therefore, denying the results of the exchange rate method also denies the results of the ICP method to a certain extent. In sum, the exchange rate method is not as poor as claimed by scholars. The greatest advantage of the ICP method is its structural measurement of the PPP. Therefore, implementing the ICP method can greatly promote the economic statistical levels of different countries. However, this advantage is gained at the expense of a huge input in collecting economic statistics.

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

29

2.3  Spatial Structure and Its Measure Comparison The greatest uniqueness and advantage of the ICP method over the exchange rate method is that the former is a structural analysis method. Therefore, the proper design and processing of the comparison structure problems are very important for the ICP method. Given the different price levels, participating economies, groupings, and methods, the PPPs in different benchmark years are incomparable. This calculation property shows that different comparison structures lead to various comparison conclusions. In other words, the comparison result of the ICP method is closely correlated with its comparison structure; that is, the former is highly sensitive to the latter. As the results of such correlation, the selection of different comparison structures is also important. The existing comparison structure should be verified as the best for people to believe conclusions from the international comparison. The author believes that special attention should be given to the five aspects of comparison structure, namely, scope, grouping, scale and internal structure, statistical capacity of participating economies, and selection of measures and comparison methods. 2.3.1   Scope of Participating Economies Generally, a larger number of economies participating in international comparison will lead to a wider comparison scope and more realistic global comparison results. Among all the countries that participated in the eighth round of ICP, only Argentina was absent in the Latin American and global ICPs. In this case, we need to consider the following: How would the absence of large countries in the comparison affect the statistical results and the validity of the comparison results? In this round of comparison, Brazil accounted for 56% of the economic aggregates of Latin America. Did this result take into consideration the absence of Argentina? If so, how can the estimation and adjustment be performed properly? Japan and South Korea are economic powerhouses in the Asia-Pacific region that have participated in the global comparison but were absent in the comparison of the Asia-Pacific region. If they had participated in the comparison of the Asia-Pacific region, would the comparison results change? If so, how? Given their high economic development levels, it

30  X. SONG

makes sense that Japan and South Korea should participate in OECD comparisons. However, does the absence of these countries in the economic comparison of the Asia-Pacific region affect the comparison results? Can these countries participate in the OECD and Asia-Pacific comparisons simultaneously? Can we test the results of the Asia-Pacific comparison through calculation? In addition, given the transitivity in the calculation process of the ICP method, will the inclusion of countries with low statistical capacities lower the validity of the PPP data of European and North American countries? If so, although the inclusion of more countries will in theory increase the validity of the global PPP data, the results may not improve because of the high importance of European and North American countries (their proportion of data is higher). Therefore, the inclusion of countries with low statistical capacities may do more harm than good. 2.3.2   Grouping of Participating Economies The grouping of participating economies involves the differentiation of their economic development levels. The ICP combines the region and development levels together when grouping the participating economies. As core members of the ICP, the OECD and EU countries are grouped together, and the remaining countries are grouped according to their regions. The countries in the OECD and EU group cross over Europe, America, and Asia, which indicates the absence of an absolute association between regional grouping and geographic space. Unlike the time index, the spatial index is not based on natural geography but considers the economic space. This grouping method is correlated with other comparison factors such as economy and culture. In addition, it can be easily operated and has been supported by international organizations in different continents. It played an important role in the success of the recent two rounds of the ICP. The question is, should the regional grouping consider more factors, such as economic level, scale (both population and geographic area), living and cultural habits, and geographic distance of economies? Grouping based on more factors may improve the quality of comparison, but it increases the statistical burden. In addition, some conditions are not ripe from the perspective of the working organization. Balancing the grouping also requires a careful design.

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

31

2.3.3   Scale and Internal Structure of the Participating Economies Does the scale of an economy affect the accuracy of the ICP result? We know that PPP exists across different economies. The question is, does it exist in the same economy? The comparison of different economies should consider the variances in their economic development levels. Should the variance within each economy also be considered? For example, different regions in China have extremely imbalanced economic development levels. Specifically, although Shanghai and Shenzhen have caught up with the OECD countries in terms of economic level, many regions in central and western China are still lifting themselves out of poverty. The national average price cannot be easily justified to represent the price of large economies with great regional differences. Underdeveloped market economies, insufficient flow of economic factors and commodities, and diversified expenditures enhance the variances among different regions of an economy and reduce the representativeness of the average price, thus leading to biased comparison results. This problem is not observed in large countries with developed market economies. These large countries with sufficient market competition and smooth commodity circulation do not show significant regional differences, and thus guarantee the representativeness of the national average price. In other words, a blind person who feels a large elephant may draw a rough conclusion, but he can make an accurate judgment if he feels a haystack. In this case, should the ICP be implemented by dividing large heterogeneous economies into regions? Should homogeneous regions be set in large heterogeneous economies and participate in the international comparison? If so, what should be the appropriate scope? For example, can we test the possibility that China’s Shanghai and Guangdong participate in the international comparison of the OECD group? 2.3.4   Statistical Capacity of the Participating Economies Aside from economic homogeneity, the ICP regional survey has another basic requirement: The participating economies should have strong statistical capabilities and data that are theoretically available. Otherwise, the reliability of the ICP result cannot be ensured. This requirement shows how difficult it is to promote the ICP.

32  X. SONG

Having more participating countries in the ICP can generate results that are more representative of the actual global PPP. However, these results may also reflect a greater difference in the statistical capabilities of various economies. The eighth round of the ICP is known for having the highest number of participating economies in the comparison. These economies include countries with weak economic levels and statistical foundations. In general, the economies that have participated in the ICP in these later stages have lower economic development levels and statistical capabilities. Therefore, including more countries will increase the difficulty of ensuring the quality and the accuracy of ICP. This is a paradoxical situation. How can the adoption of the data reported by these countries affect the ICP results? If the economic statistical foundation of these countries is ignored, the expansion of ICP will fail. However, by considering such a statistical foundation, the ICP makes a compromise to a certain extent. The ICP team has adopted a series of measures to ensure comparison quality. The question is, can these measures make up for the low statistical capabilities of developing economies? Users of the ICP result should obtain the necessary statistical reports to clarify this issue. 2.3.5   Selection of Measures and Comparison Method for Different Economic Categories The national economic comparison method is a huge system that includes various measures and comparison methods. Different methods have varying characteristics and applicable situations. Therefore, it is possible for participating countries to select different processing methods, and at the same time maintain the consistency of the system. The selection of the comparison methods may greatly influence the comparison results. If developed countries adopt advanced methods and developing countries adopt rough methods, what will the self-consistency of the overall comparison look like? How does it affect the comparison result? By contrast, can consistency be improved if all countries adopt the same method? Such considerations indicate that improving the comparison methods faces a dilemma: If developing countries do not improve their methods, the overall comparison results will not improve; however, if some countries try the improved methods and most countries still adopt

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

33

the old methods, the self-consistency of the eighth round of comparison will not be easily ensured. An outstanding example is the assessment of education. The EU and OECD countries have changed their methods from an inputbased approach (indirect method) to an outcome-oriented one (direct method), while other countries still use the input-based method. How does this shift affect the precision of the education assessment of the whole system? In addition, measuring educational productivity with the Program for International Student Assessment (PISA) is also problematic. Different societies and social groups may have different perceptions of the functions of education. Is the purpose of education improving life quality or securing employment? The so-called efficiency and its measures also vary under different orientations of educational outcome. What does the PISA plan to measure and what can it measure? Has this program been recognized by global economies? The goal of education may vary in different countries and development stages and also change with public needs. Therefore, how to develop a unified measurement orientation and structure? If PISA is the best choice for adjusting educational productivity, how can the implementation condition be fulfilled? How can we adjust the educational productivity of those economies that have not carried out this project? As PISA does not correspond to all levels of education, how can we adjust the educational productivity of the levels that are not covered by PISA? Given that higher education and scientific research cannot be clearly differentiated, the productivity adjustment of the former requires systematic consideration.

2.4   Problems that Need Further Research in International Economic Comparison and Other Suggestions 2.4.1   Cautious Use of the ICP Conclusion The methodology used in the seventh round of the ICP was considered the most viable option at the time. However, such a method underwent major improvements to address several problems before conducting the eighth round. For example, when using the importance weight, 3:1 was taken as the importance index and other weights (i.e., 2:1, 5:1, and

34  X. SONG

10:1) were tested simultaneously. This practice shows the experimental feature of the ICP methodology. Therefore, the ICP is a developing economic statistical method that is open to challenge and discussion. ICP involves many artificial settings and calculation processes. Given the paradox of neutrality in econometrics, a model cannot be built completely objectively. Different experts hold various preferences for indices and models. Therefore, model building is deeply affected by the methodological preferences of the constructor. This influence should be considered when evaluating the international comparison results. The World Bank proposes that the PPP is merely a statistical estimation. Since all statistics have errors in their abstractness, measurement, and classification, the PPP should be considered as the approximation of true value. Given the complex data collection and the PPP calculation processes, the error range cannot even be directly estimated. For these reasons, the ICP conclusion should be used with caution. 2.4.2   Problems Requiring Further Research International economic comparison is a very complicated economic statistical project. When discussing different comparison methods, we should always be aware of the problems at hand. ICP focuses on comparing the purchasing power of different currencies. However, the functions of a currency are multi-faceted and mutually affected. For example, should we measure the saving power of a currency? What is the relation of saving power with purchasing power? Does the measure of a currency’s saving power affect that of the purchasing power? Should this relation be considered in the ICP design? If so, how? As the ICP is a project that is based on GDP data, various statistical properties of GDP, including both advantages and disadvantages, are translated into the ICP. GDP has been used as the core indicator in the SNA since 1933. However, while GDP may reflect the overall development of a nation; it is expenditure that reflects the life of everyday people. What can be spent is the gross national income (GNI), which should be used to decompose the expenditure. In this sense, should the ICP be performed using GNI? The ICP can adopt multiple integrated methods to calculate the expenditure price ratio. What does this mean for the overall comparison result? Do the ICP results have multiple possibilities? If so, how should we explain the economic significance of the results?

2  RE-EXAMINING SOME PROBLEMS IN THE METHODOLOGY … 

35

2.4.3   Supplementary Suggestions for Further Expanding International Comparison Research We propose three supplementary suggestions in addition to the above-mentioned improvements. First, a PPP homogeneity index should be designed. Comparing the CPI representative basket of goods of a country with its ICP representative basket of goods can strictly reflect the similarities and differences of its consumption structure and obtain the core homogeneity ratio. Next, the homogeneity increment index can be obtained by decomposing the expenditure outside the ICP representative basket of goods. The sum of the core and the increment indices can generate the total homogeneity index, which can be used to examine the representativeness of the PPP. Second, special studies on international comparison should be conducted. Instead of rejecting the opinions of some economies on the ICP results, the global multilateral comparison results should be verified with the bilateral comparison results, and important economies (economies with great internal development difference and economies with large scale) should be mobilized and united to perform the analysis. In addition, the results of other investigations, such as the investigation of global cost of living by Mercer Consulting in the United States, can be borrowed to estimate the reliability of the global comparison conclusion. Special research should also be conducted on the comparison result of emerging economies. This study should focus on large regions with sizable populations, large GDP scale but low level of market development. For example, Brazil, Indonesia, India, and China may also be selected as the target economies in the research to further support the conclusions of global comparison. Urban economies (metropolitan economies), such as Singapore, Hong Kong, and Luxembourg, should also be investigated to determine the correlation between the results of the ICP and exchange rate methods and the difference in international comparison between urban economies and other economies. Third, open development strategies for the ICP should be adopted and made available to future societies. These strategies may include considering in advance the influence of big data on ICP, discussing the prospects of calculating basic PPP without using representative goods in deduction, and exploring the contribution of big data on the different stages of ICP development.

36  X. SONG

The ICP-related research should also be open to criticism. Many scholars and practitioners in the field of economics have different opinions about ICP. Some economists even conclude that the ICP has no future. These opinions should be collected, the related problems analyzed, and questions answered. In particular, the traditional inclination of promoting the ICP method and rejecting the exchange rate method should be discarded, and parallel development of both methods should be encouraged. The results from both methods should be compared and utilized rather than only recognizing the results from one method.

CHAPTER 3

Limitations of the Exchange Rate Method

3.1  The Exchange Rate and PPP Methods All countries have their own statistical systems and publish their GDPs every year. Much useful statistical information about a country, such as its economic growth, structural change, factor mobility, and foreign economic and trade collaboration, can be obtained as long as the statistical method remains consistent. However, horizontal comparison across countries is more complicated. Generally, no international organization or foreign institution can calculate a country’s data on its behalf. The international horizontal comparison can only use the GDP data, which use the local currency as its unit. A currency must be selected as the benchmark for comparison, and the GDPs of countries that are expressed in local currency should convert to benchmark currency before performing a horizontal comparison. The exchange rate and the PPP methods are GDP comparison techniques commonly employed in many countries. The exchange rate method divides the GDPs of countries that are expressed in the local currency by the exchange rate and then converts the results into the benchmark currency (e.g., US dollar) for comparison. For example, in 2014, the exchange rate of RMB against the US dollar was 1:6.2. People commonly use the exchange rate to translate two currencies for trade settlement, traveling abroad, and foreign investment. The exchange rate method is very simple and practical. In international © The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_3

37

38  X. SONG

trade and finance, converting different currencies using the exchange rate method has become a matter of course. However, the adoption of the exchange rate method in comparing the economic scales of two or more countries (by using GDP) has led to confusion. Therefore, people are constantly improving the PPP method as an alternative to the exchange rate method. The PPP method follows a three-step calculation process. First, a benchmark currency (e.g., US dollar or Hong Kong dollar) is selected for comparison. Second, a purchasing power conversion factor is calculated using the relative price between the commodities and services of countries (PPP). Third, the nominal GDP that is expressed in the local currency is divided by the conversion factor to obtain the GDP that is expressed in the benchmark currency. The main difference between the PPP method and the exchange rate method lies in their conversion factors. When the GDPs of the two economies are calculated according to the country price level and expressed in the local currency, the ratio of two GDPs includes the following component ratios:

GDP ratio = price level ratio × quantity ratio × currency ratio

(3.1)

The GDP ratio of the exchange rate method (GDPXR) can be obtained by dividing both sides of Eq. (3.1) by the currency ratio (exchange rate), that is, by converting the GDP ratio into one expressed by the unified currency. This ratio contains the following components:

GDPXR ratio = price level ratio × quantity ratio

(3.2)

The PPP, which is defined as the space price deflator and the currency convertor, has the following components:

PPP = price level ratio × currency ratio

(3.3)

When adopting PPP, the GDP PPP ratio obtained by dividing Eqs. (3.1) by (3.3) only has the quantity ratio:

GDP PPP ratio = GDP ratio/PPP = Quantity ratio. The above equations show that in international comparison, the GDP that is expressed in unified currency after conversion with the exchange rate method not only reflects the differences between the commodities and the services produced by different economies but also the differences in their price levels. Given the different price levels across countries, the

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

39

exchange rate method cannot objectively evaluate the variance in the quantity of their commodities and services. After converting the GDPs that are expressed in the unified currency using the PPP method, the evaluation can be performed according to the unified price level to accurately reflect the quantity difference between the two economies. The exchange rate method was widely applied in the international comparison of GDPs before the PPP method became popular. However, owing to the flaws in the exchange rate method, it has been gradually replaced by the PPP method. People often prefer a new method because they subjectively consider it as more scientific and closer to the true value. However, what is the true value? If different true values can be obtained from different observation angles, then perhaps more than one true value can exist. Both the PPP method and the exchange rate method have their own strengths, that is, they measure and reflect different aspects of the GDP comparison. Therefore, the differences between these methods must be scrutinized.

3.2  Inherent Contradictions of the Exchange Rate Method Using the exchange rate method to calculate GDP has a long history. The raw GDP statistics of a country should be expressed in the local currency, e.g., US dollar for the United States, pound for the United Kingdom, and CNY for China. A benchmark currency should also be used as a foundation in horizontal comparisons. As the US dollar is both a country currency and an international reserve currency, people multiply the GDPs of their countries by the USD exchange rate before analyzing the data and making comparisons. When writing a comparison report of the GDP of Asian countries, the Asian Development Bank uses the Hong Kong dollar as the base currency for comparison. As this method is simple and practical and does not involve any additional work in the conversion process, people have been accustomed to using the exchange rate method. Many media and politicians also prefer this simple conversion method because of their limited statistical knowledge. The formula of the exchange rate method is expressed as follows:  PY GDP = Ex

40  X. SONG

where Ex is the exchange rate, P is the price, Y is the quantity of commodities or services, and is the sum. To calculate the GDP, the quantities of the commodities are multiplied by the domestic price, and then, the results are summed up and multiplied by the exchange rate. The formula of the exchange rate method contains three elements. Unfortunately, these elements all bring contradictions that are difficult to solve in the horizontal comparison of the GDPs of different countries. First, price difference leads to the deviation of statistics. The raw GDP statistics must adopt the domestic price. The purchasing power of one US dollar may differ from one place to another as prices greatly vary across countries. For example, the price of millet in the international market is several times higher than that in Shaanxi. However, farmers in Shaanxi are unaware that millet is expensive in North America. If the price of millet is calculated according to the price in Shaanxi, the output will be much lower than that computed according to the price in North America. Economies with high income have high price levels, whereas economies with lower income have lower price levels. The difference in price levels of non-trading products is greater than that of trading products between countries with high and low incomes. Before imposing additional tariffs, subsidies, and trade costs, the prices of trading products are determined by the law of one price, while those of non-trading products are determined by the local conditions (especially wages). Economies with high income usually have higher wages and also higher costs for services. If the difference in the price levels of non-trading products is not considered when converting the GDP into a unified currency, the scales of economies with high income and price levels will be exaggerated, while those of economies with lower income and price levels will be underestimated. It is almost impossible to eliminate the regional price difference. In the market economy, the price level reflects the scarcity of resources. Therefore, rare items are given a higher value than common ones. As China has abundant labor resources, the wage in China is lower. As France has a rich supply of grapes, the price of wine in France is cheaper. The uneven distribution of resources across countries increases the trading cost of some products or decreases the tradability of some products in the market. Therefore, price differences across countries become inevitable.

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

41

Second, many commodities in developing countries have low degrees of monetization and marketization. Some of these commodities are untradeable and are produced and consumed domestically. For example, farmers in Jiangxi and Hunan brew rice wine which is mostly purchased by the local people. The local price of the rice wine is extremely low and generates little revenue. Japan also produces similar rice wine, the price of which becomes very high after commodity marketization. If the prices of these wines are converted using the exchange rate method, will several jin1 of Jiangxi rice wine be equal to one jin of Japanese rice wine? No accurate statistics can be produced for similar products if they are not commercialized and marketized. Third, not all products are available for international trading. Many commodities and services (e.g., hairdressing and housekeeping services) do not circulate across countries. If the GDP expressed by the domestic price is multiplied by the exchange rate to be converted to US dollars, all commodities and services are assumed to be included obscurely in international trading. Despite the opening up and reform of China in the past 30 years, many inland areas of the country remain closed while the coastal regions have a high degree of openness. Many products and services have never entered the international market. In fact, most commodities and services in less-developed countries and regions have a low internationalization degree. Therefore, the assumption that all commodities and services can participate in international trade is inconsistent with reality. Fourth, different statistical scopes can lead to distortion. Strictly speaking, all products in a country must be included when calculating its GDP. However, given that tens of thousands of different products and services exist in the world, including them all in the GDP calculation is impossible. Instead, people only select some representative products when calculating the GDP. The total outputs of all departments should be considered in the calculation. Each country has a unique economic structure. For instance, high-tech products account for a large proportion of the products of developed countries, but these products hardly exist in developing countries. People also select different statistical objects because of the differences in the national conditions of countries. In other words, the US statistical “basket of goods” contains many 1 A

Chinese unit of measurement. One jin equals 500 grams.

42  X. SONG

high-tech products, but the statistical baskets of most developing countries do not include these items. The inclusion of different products in the basket will inevitably lead to the deviation of statistics. Fifth, in the summation process, the service industry data of many developing countries are lost, thus leading to a huge deviation. The service industry data are obtained mainly through taxes. Some service areas in the United States, such as legal, medical, insurance, and education, have been highly industrialized and the proportion of revenue from income taxes in the total taxes is higher. Therefore, the GDP of the US service industry can be deduced according to the personal income tax paid by, for instance, lawyers and doctors. However, given the small proportion of personal income tax and the weak tax payment consciousness of the public in most developing countries such as China, the GDP of service industries in these countries cannot be easily computed. Moreover, many service departments (e.g., housekeeping services and small businesses) are not included in the statistical scope, which leads to great data loss. Sixth, fluctuations in the exchange rate significantly affect the GDP. The exchange rate method converts the GDP to that expressed by the unified currency and implies that the exchange rate can fully reflect the currency ratio. However, given that the supply and demand of currency is affected by currency speculation, interest rates, government intervention, and capital flows between economies, the exchange rate cannot fully reflect the currency ratio. Exchange rate fluctuations seriously distort the economic growth trend. Sometimes, the exchange rate fluctuation leads to a ridiculous level of GDP distortion. The GDP of Japan, which is converted according to the exchange rate method, presents an ideal example (see Table 3.1 and Fig. 3.1). Before 1970, the Japanese yen was strictly pegged to the US dollar. The GDP growth rates remained almost the same regardless of whether the yen or the US dollar was taken as basis. Given the f­ree-floating nature of the yen, the GDP growth rate of Japan with the yen as basis greatly deviates from that with the US dollar as basis. In 1978, the nominal growth rate of Japan’s GDP with the yen as basis was 10.1%, while that with the US dollar as basis was 40.5%. The latter was four times larger than the former. Given a sharp increase in the yen during the mid- and late 1980s, the nominal growth rate of Japan’s GDP with the US dollar as basis increased to 48.1, 21.2, and 21.3% in 1986, 1987, and 1988, respectively. This high-speed economic growth is actually a

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

43

Table 3.1  Effects of exchange rate fluctuations on the estimates of Japanese GDP Year

GDP (billion yen, constant price)

GDP (billion yen, current price)

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

72,176 80,869 88,074 95,537 106,692 112,901 124,913 138,756 156,631 176,175 174,371 182,565 197,925 213,824 211,204 217,733 226,388 236,327 248,786 262,430 269,824 281,094 290,586 299,480 312,848 332,662 342,080 356,131 381,582 402,074 424,479 438,590 442,182 442,939 446,764 455,442 467,329 474,786

15,951 19,263 21,860 25,019 29,430 32,742 38,026 44,561 52,776 61,994 75,265 82,814 94,814 115,444 137,759 152,211 170,935 190,482 209,756 227,347 246,465 264,966 278,179 289,315 307,499 330,261 345,644 359,458 386,428 416,246 449,392 476,431 487,961 490,934 495,743 501,707 511,935 523,198

GDP (billion USD, current price) 44 54 61 69 82 91 106 124 147 172 209 236 313 425 472 513 576 709 997 1037 1087 1201 1117 1218 1295 1385 2051 2485 3015 3017 3104 3537 3853 4415 4850 5334 4706 4324

GDP growth rate at constant price (%) – 12.0 8.9 8.5 11.7 5.8 10.6 11.1 12.9 12.5 −1.0 4.7 8.4 8.0 −1.2 3.1 4.0 4.4 5.3 5.5 2.8 4.2 3.4 3.1 4.5 6.3 2.8 4.1 7.1 5.4 5.6 3.3 0.8 0.2 0.9 1.9 2.6 1.6

GDP growth rate at current JPY price (%) – 20.8 13.5 14.5 17.6 11.3 16.1 17.2 18.4 17.5 21.4 10.0 14.5 21.8 19.3 10.5 12.3 11.4 10.1 8.4 8.4 7.5 5.0 4.0 6.3 7.4 4.7 4.0 7.5 7.7 8.0 6.0 2.4 0.6 1.0 1.2 2.0 2.2

GDP growth rate at current dollar price (%) – 20.8 13.5 14.5 17.6 11.3 16.1 17.2 18.4 17.5 21.4 13.0 32.4 35.9 11.0 8.7 12.4 23.1 40.5 4.1 4.8 10.5 −7.0 9.1 6.3 6.9 48.1 21.2 21.3 0.1 2.9 14.0 8.9 14.6 9.9 10.0 −11.8 −8.1 (continued)

44  X. SONG Table 3.1  (continued) Year

GDP (billion yen, constant price)

GDP (billion yen, current price)

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

465,275 464,348 474,830 476,518 477,898 485,951 497,423 503,903 512,434 523,667 518,212 489,571 511,284 507,705

512,439 504,903 509,860 505,543 499,147 498,855 503,725 503,903 506,687 512,975 501,209 471,139 481,773 468,258

GDP (billion USD, current price) 3915 4433 4731 4160 3981 4303 4656 4572 4357 4356 4849 5035 5488 5867

GDP growth rate at constant price (%) −2.0 −0.2 2.3 0.4 0.3 1.7 2.4 1.3 1.7 2.2 −1.0 −5.5 4.4 −0.7

GDP growth rate at current JPY price (%) −2.1 −1.5 1.0 −0.8 −1.3 −0.1 1.0 0.0 0.6 1.2 −2.3 −6.0 2.3 −2.8

GDP growth rate at current dollar price (%) −9.5 13.2 6.7 −12.1 −4.3 8.1 8.2 −1.8 −4.7 0.0 11.3 3.8 9.0 6.9

Data Source World Development Indicator (WDI) 2012 ed. World Bank. http://data.worldbank.org/ data-catalog/world-development-indicators

Fig. 3.1  Japanese GDP growth rates at different currencies between 1960 and 2011 (Data Source WDI, 2012 ed. World Bank)

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

45

false image caused by the exchange rate adjustment. The real economic growth rate of Japan was only 2.8, 4.1, and 7.1% in those years. During the Asian financial crisis in 1997, Japan’s economic growth with the US dollar as basis dropped by 8.1 and 9.5% in 1997 and 1998, respectively, because of the rapid devaluation of the yen. This decrease exaggerated the influence of the financial crisis on the Japanese economy from the opposite direction. However, the actual GDP growth rate of Japan during the financial crisis decreased only by 2.0% in 1998 and gradually recovered in 2000. Interestingly, the US dollar was down against the yen during the US subprime mortgage crisis, which evolved into the global financial crisis at the end of 2007. As a result, the GDP growth rate of Japan with the US dollar as basis increased to 11.3% in 2008, but the actual GDP growth rate was only –1.0% during that year. Therefore, the exchange rate fluctuation can greatly disturb the international comparison of GDPs. When the local currency or the US dollar demonstrates wide fluctuations, the changing trend of the GDP converted by the exchange rate method becomes very different from reality. The economic growth rate of China presents another persuasive example. In 1980, the exchange rate between the US dollar and RMB was 1:1.70, which increased to 1:8.62 in 1994 after the RMB exchange rate began to depreciate. Meanwhile, the Chinese economy began to take off with an average annual economic growth rate of over 9.0%. However, the exchange rate method rejects this economic miracle of China (see Table 3.2). In 1994, China implemented the exchange rate reform by combining RMB with the foreign exchange certificate under the dual-track system. As a result, the nominal exchange rate increased from 1:5.76 to 1:8.62. If calculated according to the exchange rate method, the economic growth rate of China was –8.8% in 1994. This figure may give the impression that China suffered significant setbacks in 1994. However, the actual nominal GDP growth rate of China at that time was 13.1%. The RMB continued to appreciate after 2005. If converted directly using the exchange rate between the US dollar and RMB, the economic growth rate of China was as high as 29.4% in 2008, when the actual growth rate was only 9.6%. In 2013, the actual GDP growth rate of China was 7.4%, but this figure was 11.6% when computed on the basis of the US dollar converted using the exchange rate method. The results from the exchange rate method are higher because of the appreciation.

46  X. SONG Table 3.2  Economic growth rates of China Years

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

GDP (100 million CNY) 18,668 21,782 26,923 35,334 48,198 60,794 71,177 78,973 84,402 89,677 99,215 109,655 120,333 135,823 159,878 184,937 216,314 265,810 314,045 340,903 401,202 471,564 519,470 568,845

Nominal economic growth rate (%) 3.8 9.2 14.2 14.0 13.1 10.9 10.0 9.3 7.8 7.6 8.4 8.3 9.1 10.0 10.1 11.3 12.7 14.2 9.6 9.2 10.4 9.2 8.4 7.4

Exchange rate Economic growth (CNY/1 USD) rate by the exchange rate method (%) 4.78 5.32 5.51 5.76 8.62 8.35 8.31 8.29 8.28 8.28 8.28 8.28 8.28 8.28 8.28 8.19 7.97 7.60 6.95 6.83 6.77 6.46 6.31 6.19

−14.9 4.8 19.3 25.6 −8.8 30.2 17.6 11.3 7.0 6.3 10.6 10.5 9.7 12.9 17.7 16.9 20.2 28.8 29.4 10.4 18.8 23.1 12.8 11.6

Data Source China Statistical Abstract, 2014; and the China Economic Database

Clearly, the exchange rate cannot accurately reflect the changing trend of the macroeconomy in the case of exchange rate fluctuation. These two examples—the appreciation of Japanese yen and the depreciation–appreciation of the Chinese yuan—demonstrate that the exchange rate must be relatively stable when calculating the GDP using the exchange rate method. If the exchange rate undergoes huge fluctuations, the GDP tends to be overestimated or underestimated, thus leading to confusion about the status of the economy. To prevent such fluctuations from generating meaningless changes in the international economic statistical indicators, the World Bank uses the average value of three-year exchange rates to ease the exchange rate

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

47

fluctuations and to eliminate the disturbance that results from short-term exchange rate fluctuations. However, the three-year span is too short to eliminate the influence of the significant appreciation and depreciation of exchange rates.

3.3  Origin of Official GDP Data by the Exchange Rate Method In international comparison, the exchange rate method uses the exchange rate to convert the economic indicators calculated with different monetary units and to maintain the comparability of the indicators of different countries. This method does not directly adopt the market exchange rate or the official exchange rate, but it guarantees smoothing to eliminate the influence of short-term or accidental factors and to reflect the long-term trends of the purchasing power of different currencies in the international market. The commonly used exchange rate adjustment methods include the ATLAS method of the World Bank and the PARE of the UN. The ATLAS method focuses on the simple arithmetic mean of the average market exchange rate in the present year and the average market exchange rate after adjusting the price of the previous two years. This method eliminates the short-term fluctuation in the exchange rates between a specific currency and the US dollar through price adjustment. The ATLAS method has better stability than the market exchange rate. Using this coefficient as the currency conversion factor for GDP comparison, that is, using the comparable international price to measure the actual economic outputs of different countries, can accurately measure the gap in the economic development of different countries. The ATLAS method has been adopted by the World Bank since 1994 to calculate the real per capita incomes of countries. The results of the ATLAS method are then used as the standard to classify countries into high-, medium-, and low-income levels. PARE method of the UN produces a currency conversion coefficient by taking the average market exchange rate within a certain period or year as the benchmark exchange rate, adjusting the benchmark exchange rate with the price indices of countries (GDP deflator in general), and then extrapolating the benchmark to other periods.

48  X. SONG

Although the averaging of the ATLAS method aims to eliminate the short-term exchange rate fluctuations, this method only weakens the effect of these fluctuations to a certain extent because of its limited coverage period (three years). In this case, fluctuations beyond those three years, especially the significant exchange rate appreciation and depreciation caused by non-economic factors, cannot be removed. Therefore, this method only makes running repairs of the exchange rate. This method may also yield confusing results in cases when the exchange rate rapidly changes and the inflation rate increases to a high level. The PARE method is a currency conversion method that converts the GDP index at a constant price according to the unified exchange rate. This method assumes that the benchmark exchange rate is close to or truly reflects the parity relationship between the commodities and the services of the countries in comparison. The results of this method are affected by the benchmark exchange rate instead of the changes in the exchange rate in different years. This method also eliminates the influence of price fluctuation in a certain period. The GDP after conversion is comparable over time. Comparatively speaking, the converted GDP can more accurately reflect the economic development gap among countries. The key element of this method is whether the assumption of the benchmark exchange rate is close to the truth. In addition, the extrapolated years must not be too far from the year of the benchmark exchange rate. The ATLAS and PARE methods compare the actual GDPs of various countries from different perspectives. The former performs a comparison according to the present price level, and the GDPs converted according to this coefficient are incomparable in time. The latter performs a comparison according to a constant price, and the GDPs converted according to this coefficient are comparable over time. In sum, the PARE method places the economic aggregates of a country at the same level of world economic development for measurement and comparison, and it considers both the quantity and the quality of economic development. Moreover, the PARE method not only avoids the influence of factors such as government intervention on the data but also provides timely comparison results because of its easy and simple operation. The GDPs calculated by the exchange rate method and published by the World Bank include both GDPs at current prices and GDPs at constant prices, with 2005 as the base period. These GDPs are computed with the ATLAS and PARE methods, respectively. Tables 3.3 and 3.4 present the GDPs at the current and constant (2005 base year) prices of

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

Table 3.3  GDP at current price for China, the United States, and Japan from 1960 to 2013 using the exchange rate method (Unit: billion USD)

49

Year

China

United States

Japan

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

59.18 49.56 46.69 50.10 59.06 69.71 75.88 72.06 69.99 78.72 91.51 98.56 112.16 136.77 142.25 161.16 151.63 172.35 148.18 176.63 189.40 194.11 203.18 228.45 257.43 306.67 297.83 270.37 309.52 343.97 356.94 379.47 422.66 440.50 559.22 728.01 856.08 952.65 1019.46 1083.28 1198.47 1324.81

520.53 539.05 579.75 611.67 656.91 712.08 780.76 825.06 901.46 973.39 1075.90 1167.80 1282.40 1428.50 1548.80 1688.90 1877.60 2086.00 2356.60 2632.10 2862.50 3210.90 3345.00 3638.10 4040.70 4346.70 4590.10 4870.20 5252.60 5657.70 5979.60 6174.00 6539.30 6878.70 7308.70 7664.00 8100.20 8608.50 9089.10 9665.70 10,289.70 10,625.30

44.31 53.51 60.72 69.50 81.75 90.95 105.63 123.78 146.60 172.20 209.07 236.15 312.74 424.89 471.64 512.86 576.41 709.40 996.74 1037.45 1086.99 1201.47 1116.84 1218.11 1294.61 1384.53 2051.06 2485.24 3015.39 3017.05 3103.70 3536.80 3852.79 4414.96 4850.35 5333.93 4706.19 4324.28 3914.57 4432.60 4731.20 4159.86 (continued)

50  X. SONG Table 3.3 (continued)

Year

China

United States

Japan

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

1453.83 1640.96 1931.64 2256.90 2712.95 3494.06 4521.83 4990.23 5930.50 7321.89 8229.49 9240.27

10,980.20 11,512.20 12,277.00 13,095.40 13,857.90 14,480.30 14,720.30 14,417.90 14,958.30 15,533.80 16,244.60 16,800.00

3980.82 4302.94 4655.80 4571.87 4356.75 4356.35 4849.18 5035.14 5495.39 5905.63 5937.77 4901.53

Data Source World Bank Database. http://databank.worldbank.org/ data/views/reports/tableview.aspx

Table 3.4  GDP at constant price for China, the United States, and Japan from 1960 to 2013 using the exchange rate method (Unit: billion USD)

Year 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

China 87.99 87.73 63.78 60.21 66.35 78.49 91.83 101.66 95.86 91.93 107.47 128.32 137.30 142.52 153.78 157.32 171.00 168.27 181.05 202.18 217.49 234.55 246.85 269.20 298.42

United States 2794.81 2859.09 3033.50 3166.97 3350.66 3565.10 3796.83 3891.75 4078.55 4204.99 4339.84 4482.70 4718.01 4984.23 4958.47 4948.63 5215.22 5455.59 5758.97 5941.85 5927.32 6081.12 5964.94 6241.28 6694.33

Japan 654.85 733.72 799.09 866.80 968.01 1024.35 1133.32 1258.92 1421.10 1598.42 1582.06 1656.40 1795.76 1940.01 1916.24 1975.48 2054.00 2144.18 2257.22 2381.01 2448.10 2550.35 2636.46 2717.16 2838.45 (continued)

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

Table 3.4 (continued)

51

Year

China

United States

Japan

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

343.71 389.99 424.49 473.66 527.10 548.51 569.57 621.85 710.41 809.61 915.51 1015.53 1117.17 1221.04 1316.69 1417.01 1534.63 1674.01 1841.83 2027.58 2256.90 2543.00 2903.15 3182.86 3476.13 3839.28 4196.33 4517.46 4864.00

6978.12 7223.19 7473.22 7787.36 8074.01 8228.92 8222.94 8515.28 8749.03 9102.18 9349.64 9704.54 10,140.02 10,591.23 11,104.54 11,558.79 11,668.44 11,875.70 12,207.14 12,670.77 13,095.40 13,444.60 13,685.24 13,645.50 13,263.10 13,595.64 13,846.78 14,231.57 14,498.62

3018.22 3103.67 3231.15 3462.07 3647.99 3851.27 3979.30 4011.89 4018.75 4053.46 4132.19 4240.04 4307.70 4221.41 4212.99 4308.10 4323.41 4335.93 4409.00 4513.08 4571.87 4649.26 4751.18 4701.69 4441.83 4648.47 4627.42 4694.39 4766.66

Data Source World Bank Database, http://databank.worldbank.org/ data/views/reports/tableview.aspx

China, the United States, and Japan between 1960 and 2013 computed using the exchange rate method. As shown in the tables, the United States is far ahead of the two other countries in terms of GDPs at both current and constant prices. However, the gap between China and the United States has narrowed. In terms of GDP at current prices, China was surpassed by Japan in 1961 but surpassed Japan in 2010. In terms of GDP at constant prices, the GDP of China was less than 1/8 of that of Japan in 1961 but was larger than that of Japan in 2013. On the one hand, these results prove that the exchange rate can accurately reflect the trend of GDP. On the other hand, they also show that different conclusions may occur when comparison is conducted only in specific years.

52  X. SONG

3.4   Premises of Using the Exchange Rate Method Adopting the exchange rate method to compare the GDPs of different countries has several premises. First, a highly developed common market must be established among the countries in comparison. Second, the price difference among countries must be controlled through market trading. Third, the statistical calibers of the GDPs of different countries must be consistent. Fourth, the exchange rates of comparison countries must be stable. Given that Western countries closely satisfy the above conditions, their economic data can be compared with one another using the exchange rate method. The GDP ranking of seven industrial countries can reflect their relative economic scales. Given that China is still transforming from a planned economy to a market economy, some of its statistical calibers are different from those of other countries. Even if some statistical items have the same names, their connotations may differ. The exchange rate of China constantly adjusts with wide fluctuations; therefore, adopting the exchange rate method to estimate the GDP of China will inevitably lead to distortions. These reasons are why many people misunderstand the economic scale and the foreign trade situation of China. Both the primary strength and weakness of the exchange rate method lie in its simple nature. The exchange rate method emphasizes the comparison of the total output value between two (or more) countries. The value of a commodity or service includes three elements, namely quantity, quality, and price. Examining the national economic scale focuses on how many commodities (or services) that two (or more) countries actually have instead of how much these commodities are worth in the domestic market. If these commodities have a high price, their total output value must also be high. However, the total volume of these commodities is not necessarily high. The discussion on people’s living standards mainly focuses on how many commodities and services each person can consume. Higher prices generate higher values, but it does not mean the public is more satisfied. If prices continue to increase sharply, people tend to spend more money at the expense of their life quality. When evaluating the GDP, people pay more attention to their standard of living than trade. The amount of nominal income does not matter; how many commodities can be bought by this amount is more important. If the GDP of country

3  LIMITATIONS OF THE EXCHANGE RATE METHOD 

53

A is greater than that of country B, and people in country A must spend twice as much on the same commodity as country B because of higher price levels, then people in both countries enjoy the same material conditions and living standards. For example, McDonald’s hamburgers sold in New York and Tokyo share the same quality. The dollar amount people spend on one McDonald’s hamburger in New York can buy three McDonald’s hamburgers in Tokyo. In other words, the purchasing power of one US dollar in Tokyo is three times that of one US dollar in New York. However, for tourists, we cannot say that with the same level of spending eating one hamburger in New York is equivalent to eating three hamburgers in Tokyo. Under the PPP method, the value of a hamburger in New York must be equal to its value in Tokyo. The value can be obtained by multiplying the quantity of hamburgers by their price in New York.

CHAPTER 4

Viewing and Applying the PPP Correctly

4.1  Currency Purchasing Power and PPP 4.1.1   Starting Point of the Horizontal Comparison Currency purchasing power refers to the quantity of commodities and services that can be purchased by a certain unit of money. Purchasing power parity (PPP) refers to the ratio of the purchasing power of two countries’ currencies for a certain amount of commodities and services, that is, the price ratios of two currencies for buying the same quantity and quality of commodities. In most cases, when evaluating GDP, the economic strength of a country is not assessed by the number resulting from the simple conversion of the GDP into the US dollar amount. Instead, the actual output of the country is more important. Moreover, nominal income is not important when discussing their living standards. The difference between the wealthy and the poor depends on how many products and services can be consumed. If the GDP of country A is twice as high as that of country B but people in country A need to spend twice as much on a product than those in country B because of high prices, the actual GDPs and material conditions of the residents in these countries will be the same. Under the PPP method, if a hamburger costs $4 in New York, this commodity must be similarly priced in Tokyo, Paris, or other places. Multiplying the quantity of hamburgers in Tokyo and Paris by the price of hamburgers in New York yields the value that is calculated according © The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_4

55

56  X. SONG

to the PPP. In this way, the outputs between two countries can be compared fairly. If a basket of commodities of the same quantity and quality cost 60 CNY in China or $10 in the United States, then the PPP of RMB against the US dollar is 6:1. In other words, spending $1 on these commodities is equal to spending 6 CNY. People often compare the PPPs of different currencies using readily available examples. However, performing an international comparison with the purchasing power of a commodity can lead to a paradox. For example, if a McDonald’s hamburger is sold at 30 CNY in China and $4.5 in the United States, the purchasing power of $1 is appropriately equal to that of 6.6 CNY, which is close to the USD/CNY exchange rate of 1:6.2. Assume that a haircut in New York costs $30 and 20 CNY in Beijing. In this case, the purchasing power of $1 is less than that of 1 CNY. When a Louis Vuitton bag is sold at $200 in the United States and 4000 CNY in China, the purchasing power of $1 is almost equal to that of 20 CNY. In sum, the relative prices of a haircut and luxury commodities deviate far from the exchange rate. Many other similar examples can be found. Therefore, a scientific and rigorous method that can comprehensively evaluate the purchasing power of a currency must be devised, rather than evaluate the relationship between two currencies according to the relative price of a commodity or service. 4.1.2   Absolute PPP Swedish economist Gustav Cassel expounded the PPP theory in 1922 and argued that under the equilibrium condition, the relative purchasing power of two currencies should be reflected in their relative values (Cassel 1922). PPP is not only important in international macroeconomics and traditional exchange rate determination theory but has also become the theoretical basis of the ICP. In economics, PPP is categorized into absolute and relative PPP. The theoretical foundation of absolute PPP is the law of one price, which means that one commodity must be sold at the same price in different countries within a fully competitive market without transportation cost and trade barriers. This law affirms the validity of the foreign exchange market and holds that arbitrage space can disappear within a very short period. According to the law of one price, the absolute PPP can be expressed as follows:

E = P0 /Pi

4  VIEWING AND APPLYING THE PPP CORRECTLY 







57



where P0 = P10 , P20 , · · · , Pn0 , Pi = P1i , P2i , · · · , Pni , i = 1, · · · , k. Superscript 0 represents the countries or regions that are the basis for comparison. For example, the World Bank selects the United States as the base for comparison, whereas the Asian Development Bank (ADB) selects Hong Kong. Superscript i represents the countries or regions participating in the comparison. A total of k countries or regions have participated in comparison. The subscript represents the commodities or services participating in the comparison. A total of n commodities or services are considered. P0 and Pi represent the price vectors of base countries (regions) and countries (regions) in comparison, respectively. When calculating the domestic and foreign price levels, the commodity baskets and the weights of their commodities should be the same. Absolute PPP theory holds that the equilibrium exchange rate is determined by the purchasing power of two countries’ currencies. Clearly, the above conditions cannot be met in reality. Dornbusch (1976) proposes that trade barrier and information asymmetry not only lead to differences in the prices of commodities in various regions but also restrict arbitrage and hinder the establishment of the PPP theory. In this case, we put forward a theory of relative PPP, which enables the existence of a certain trade cost. Relative PPP theory emphasizes that the difference between the spot and the forward inflation rates must be equal to the difference in the exchange rate during the same period. Exchange rate appreciation and depreciation are determined by the difference in the inflation rates between two countries. Although relative PPP can effectively reflect the influence of the actual exchange rate fluctuations on the price, many scholars contend that this theory is still too idealistic. 4.1.3   Deviation of the Exchange Rate from PPP The exchange rate often deviates from PPP in both short and long terms. 1. Short-term deviation. The establishment of the law of one price strictly depends on international commodity arbitrage. If the com­modities are transported and traded at no cost, the prices of the same commodities as expressed by one currency will become equal in different countries. Given the non-trading commodity components in trading commodities, trade cost, and price stickiness, the law of one price cannot be established in reality. Isard

58  X. SONG

(1995) argues that the short-term establishment of PPP is usually questioned because the same-currency commodities from the fine manufacturing industries of different countries may greatly and constantly deviate from the exchange rate. According to Rogoff (1996), given the non-trading components in commodities, products that are produced all over the world (i.e., McDonald’s hamburgers) will have different prices when they are simultaneously converted to the US dollar according to the exchange rates of each country. The mobility restriction of non-trading commodities prevents the occurrence of international arbitrage, and thus the establishment of the law of one price is affected. Obstfeld and Rogoff (2009) argue that trade cost is a key factor in explaining the differences among international prices. It hinders the arbitrage between countries and results both in the non-linear adjustment of the exchange rate, and in the deviation of the exchange rate from the PPP. 2.  Long-term deviation. Although some people have attempted to prove that the exchange rate follows the PPP in the long term, other studies have rejected such a possibility. In 1964, Balassa proposed the productivity bias hypothesis, which refers to the deviation of a long-term equilibrium exchange rate from the PPP. Countries with high productivity levels have trading and non-trading commodities with high price ratios, and the price levels used as the base of the PPP in high-income countries are higher than those determined according to the exchange rate (Balassa 1964). Therefore, the exchange rate calculated according to the PPP is lower than the long-term equilibrium exchange rate. As the difference in the productivity levels between high- and low-income countries continues to increase, the PPP of low-income countries will be underestimated. Officer (1976) contradicts Balassa by arguing that the quality of non-trading commodities in high-income countries is higher than that of non-trading commodities in low-income countries. However, Balassa argues that the quality of non-trading commodities in developed countries must be as high as that of non-trading commodities in lessdeveloped countries to offset the high internal price ratio. Hsieh (1982) examines the actual exchange rates of Japan and Germany against that of the United States using data between 1954 and 1976. He finds that the productivity difference can well explain the

4  VIEWING AND APPLYING THE PPP CORRECTLY 

59

changes in the actual exchange rates of Japan and Germany against those in the actual exchange rates of the United States. However, this argument has since been rejected (Froot and Rogoff 1991; Rogoff 1996). The productivity difference can explain the international difference of the relative prices of non-trading commodities, but the relative prices of non-trading commodities cannot explain the long-term deviation of the exchange rate against the PPP (Asea and Mendoza 1994). Lee and Tang (2000) argue that a high production rate can appreciate the actual exchange rate, but such a connection is established not through the relative prices of non-trading commodities but through those of trading commodities. Nevertheless, these studies have not yet reached a conclusion, and their findings must be validated through practice. The PPP theory is based on the traditional quantity theory of money. This theory does not distinguish trading commodities from non-trading ones when determining the exchange rate, and it ignores the influence of trade cost, trade barrier, and other non-trading factors. In fact, modern exchange rate determination theory has gone far beyond the scope of the century-old PPP theory. In this sense, the PPP theory has become outdated.

4.2  Origin and Progress of ICP 4.2.1   Two Basic Methods of International Comparison To perform an international economic comparison, the GDPs expressed in the local currency must be converted to indicators expressed in the same currency. International comparison can be performed using two methods, namely the exchange-rate-based ATLAS method of the World Bank and the PPP method. The exchange rate method has been criticized throughout its long history. Specifically, the exchange rate is affected by actual currency purchasing power, import and export commodity structures, foreign trade, international capital flow, and the foreign trade policies of different countries. GDPs that are converted using the exchange rate method may face serious distortions and are also influenced by the price level (Yuan et al. 2008). The global ICP of the World Bank aims to analyze and perform an international comparison of the economies and structures of different

60  X. SONG

countries. Specifically, this project (1) gathers the price information of several representative commodities, (2) uses the expenditure classification components of the corresponding items in the GDPs of different countries as the weights for their weighted average, (3) calculates the PPPs of different categories of commodities and services as well as the PPPs of GDPs, and (4) converts the GDPs in the local currency to the GDPs in the unified currency. The PPP method addresses many flaws of the exchange rate method in international comparison. The PPP has a more solid theoretical basis as the GDP that it converts reflects the pure quantity difference and excludes the influence of price. ICP is established on the basis of a theoretical system with strict logic and serves as a scientific framework for comparing and evaluating the actual economic scales and structures of various countries. 4.2.2   Constantly Advancing ICP In 1965, the 13th session of the UN Statistics Commission suggested the need to investigate international comparison of GDPs and build upon the experience of organizations such as the OECD and the Economic Commission for Latin America and the Caribbean. The goal is to devise a novel method for accurately evaluating and comparing the economic scales and structures of various countries. In 1968, the UN Statistics Division published the International Comparison of Production, Income, and Expenditure, which suggested that several countries should be selected to promote international horizontal comparison. Heeding this advice, the UN Statistics Commission designed the ICP, which recently completed its eighth round. More countries and regions have participated in the ICP as the theory and methodology of the project continue to mature (see Table 4.1). According to the 2011 International Comparison Program (ICP) in Asia and the Pacific, Purchasing Power Parities and Real Expenditures: A Summary Report, released by the ADB, “The International Comparison Program (ICP) is a global statistical initiative set up on the recommendation of the United Nations Statistical Commission to enable comparisons of economic aggregates. From a modest beginning with just 10 countries participating in 1970, the ICP has expanded to cover over 180 countries in the latest 2011 benchmark comparisons. The ICP, organized along regional lines, is coordinated by the ICP Global Office in the World Bank. The ADB is the Regional Coordinating Agency in

4  VIEWING AND APPLYING THE PPP CORRECTLY 

61

Table 4.1  Number of participating countries in ICP Year First round Second round Third round Fourth round Fifth round Sixth round Seventh round Eighth round

1970 1973 1975 1980 1985 1993 2005 2011

Number of participants 10 16 34 60 64 83 146 199

Data Source Asian Development Bank (2011) International Comparison Program in Asia and the Pacific. Purchasing Power Parities and Real Expenditures: A Summary Report

Asia and the Pacific region. The 2011 ICP in Asia and the Pacific follows a successful benchmark comparison in 2005 and a subsequent update of these results to 2009”.1 Twenty-three economies in Asia and the Pacific region participated in the 2011 ICP, including Bangladesh, Bhutan, Brunei Darussalam, Cambodia, China, Fiji, Hong Kong of China, India, Indonesia, the Lao People’s Democratic Republic, Macau of China, Malaysia, the Maldives, Mongolia, Myanmar, Nepal, Pakistan, the Philippines, Singapore, Sri Lanka, Taiwan of China, Thailand, and Vietnam. According to the report: “Among the important features of the 2011 ICP are the participation of Myanmar for the first time; the national coverage of the PRC price surveys compared to 11 capital cities in the 2005 ICP; and the increased coverage of the price surveys in India and Indonesia”.2 The ICP divides commodities into 155 basic headings, including over 500 core commodities and over 2000 representative commodities. To obtain the PPP of the GDPs of different countries, this project calculates the PPPs of all these categories and then uses the shares of these categories in the economic value as weights for calculating the sum of the weighted averages. Each economy collects comparable price data, compiles detailed GDP and expenditure statistics, estimates their PPPs, and provides comparable data on the GDP hierarchy and its components 1 From the Forward, 2011 ICP in Asia and the Pacific. Purchasing Power Parities and Real Expenditures: A Summary Report. P. vi. Published in 2014 by the ADB. 2 Ibid.

62  X. SONG

(measured by the same currency). The PPP method avoids the influence of the exchange rate conversion and the differences in relative prices to reflect accurately the economic strength and living standards of people within a specific economy. Given accelerating economic globalization and regional integration, adopting the PPP method has become necessary to measure and evaluate the social and economic development status of various countries. With the obvious distortions of the exchange rate method in international comparison, the contradictions among the data have become increasingly prominent, and the demand of international societies for PPP results in economic analysis, decision evaluation, and policy assessment continues to grow. People hope that the ICP results can accurately reflect world economic development and evaluate the economic and social development of different countries as well as the gap between the rich and the poor. These results provide international organizations, governments, and researchers with globally comparable economic statistics that are useful in analyzing economic and social development progress, in developing poverty reduction, in combating dumping, in creating income distribution policies, and in implementing other related decisions. 4.2.3   Increasing Maturity After Eight Rounds of Exploration Since 1968, the global ICP activities have experienced eight rounds of exploration, research, and practice. The organization and management of the project have become more rigorous, its technical methods have become more advanced, and its actual operation has been standardized. The ICP redesigned the survey framework in ICP 2011. To ensure the comparability of price survey data between countries, four unifications were implemented. First, the price collection catalog was unified. Second, the price collection scope was unified, and the design of a sampling point program in the existing CPI investigation organization system was required to ensure that the collected price data would represent the national level. Third, the price collection time frame was unified, and all countries and regions participating in the ICP were required to perform investigations in 2011 according to different investigation frequencies to ensure that the collected price data would represent the national average level for that year. Fourth, the specification of the 155 basic headings for GDP expenditure was based on the definition of the national account system from 1993.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

63

The ICP mainly investigates the prices of items in the product specification. This catalog is divided into global and regional items and involves personal consumption, government consumption, equipment, and construction. Personal consumption includes three product categories, namely products in the regional catalog, overlapping products in the regional and global catalogs, and products in the global core catalog. The price collection objectives for the government consumption expenditure and fixed capital do not differ between the global and the regional catalogs. Table 4.2 shows the distribution of the product specifications, the prices of which were collected in the eighth round of the ICP. The GDP is divided into 155 basic headings on which the ICP calculates the annual average prices of the product specifications that correspond to each basic heading to obtain the regional and global PPP through the regional and global links. The regional link is based on the regional product catalog made by regional coordination institutions and countries in the region. The product specifications in this catalog represent the consumption types of different countries and are comparable across countries. The global link is based on the global core product catalog made by the global office and the regional coordination institutions. The global core products are comparable across regions. For the basic headings, the regional PPP results are linked to the global ones using the weighted country-product-dummy (CPD) method. After calculating the PPPs of different product categories, the global and regional PPPs of different countries can be obtained through the summary and linking method, which involves the intra-regional PPP summary and interregional PPP linking methods. The former method has no uniform regulation. The EU-OECD adopts the EKS method, while other regions adopt the CPD method. The eighth round of the ICP adopts the global core product catalog method to link the regional PPP results. For the total class above the basic headings, ICP 2005 adopts the great region method to use the consumption shares of different product categories in the total economic value as weights, calculates the weighted average sum of the PPPs of different product categories, and links the results between great regions to obtain the PPPs of the GDPs of different countries stage by stage.3

3 ICP

classifies GPD into four levels. The lowest level is the basic heading.

64  X. SONG Table 4.2  Distribution of product specifications in the eighth round of ICP Category 1. Household consumption expenditure 01—Food and non-alcoholic beverages 02—Alcoholic beverages and tobacco 03—Clothing and footwear 04—Residence 05—Furnishings and household equipment 06—Health 07—Transport 08—Communication 09—Recreation and culture 10—Education 11—Restaurants and hotels 12—Other products 2. Government consumption expenditure Duty wage in government sector 3. Fixed capital formation Machinery and equipment Construction

Regional directory

Global directory

750 236 19 104 18 82 86 49 16 65 8 31 36

601 155 16 72 19 68 89 53 15 53 7 33 21

44 44 250 190 60

Data Source Asian Development Bank (2011) International Comparison Program in Asia and the Pacific. Purchasing Power Parities and Real Expenditures: A Summary Report

ICP 2011 replaced the great region method with the CAR method. Instead of selecting a benchmark in each great region, the benchmark country was left unchanged. For the GDP classification, the eighth round of ICP divides the GDP expenditures into 7 items, 26 classes, 61 groups, and 126 categories. Table 4.3 shows the main classifications of GDP expenditures and the number of items in each level of these classifications. Both the PPP theory and its associated statistical methods were developed considerably during the seventh and eighth rounds of the ICP. After eight rounds of comparison activities, the ICP method was standardized and improved through theoretical innovation and reform, and solved the technical difficulties in practice and research. Taking the eighth round of ICP as an example, ICP 2011 has demonstrated innovations in the following aspects: 1.  Develop an external communication strategy with the objective to establish close connections with data users and to increase the validity of the data service.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

65

Table 4.3  The eighth round ICP classification of expenditures Expenditure category GDP Final consumption expenditure by households Individual consumption expenditure by Non-profit Institutions Serving Households (NPISHS) Individual consumption expenditure by government Collective consumption expenditure by government Expenditure on gross fixed capital formation Changes in inventories and acquisitions, less disposals of valuables Balance of exports and imports

Category

Group

Class

Basic class

26 13

61 43

126 90

155 110

1

1

1

1

5

7

16

21

1

1

5

5

3 2

6 2

11 2

12 4

1

1

1

2

Data Source Asian Development Bank (2011) International Comparison Program in Asia and the Pacific. Purchasing Power Parities and Real Expenditures: A Summary Report

2.  Establish the ICP quality assurance framework (ICP-QAF) to improve data quality. 3.  Promote the statistical capacity-building strategy for developing countries with ICP as the motivation. 4.  Write the book Measuring the Real Size of the World Economy, which explains the framework, methods, and results of the ICP. 5.  Design the basic conceptual framework of national accounts to achieve consistency between the expenditure and price data. 6.  Introduce the price and expenditure data verification system to track the survey and collection processes in all directions. 7.  Adopt the core list method to link the global PPP results and improve the quality of the global results. This is different from the group or bridge countries method used before.4 8. Continuously improve the ICP method by determining the survey list, designing the survey framework and data collection method, and addressing the incomparability of some items (Yu 2011).

4 The 2005 ICP used the group method, which selects a group of countries, a few from each region, to provide a link between regional PPPs to form global results.

66  X. SONG

4.3  Methods of ICP 4.3.1   Two Methods of ICP The ICP mainly adopts the following methods: 1. The F method, also known as the ideal formula, is primarily used for bilateral comparison. This method uses Fisher’s geometric method5 in the index and the factor analysis method. The F method is expressed as follows: √ PPPj,n = LP   pij qin pij qij L = i , P = i p q p i in in i in qij where i = 1, 2, · · · , m, represents different commodities, n is the benchmark country, j is a comparison country, qij is the ­consumption of commodity i in country j, and pij is the average price of i in j. The F method uses the different production structure weights of two countries to obtain comprehensive currency values. However, given that the solution from the geometric average method does not reflect the unified magnitude requirement of the PPP method, the practical significance of the F method remains unclear. 2. The Geary–Khamis (GK) method is used for multilateral comparisons. This method takes a country as the benchmark and expresses the prices of other countries in their respective local currencies to obtain the international average price. The PPPj of the jth country is computed as follows:    pij qin pij qin x¯ in qij PPPj = i = i · i x ¯ q x ¯ q i i ij i in ij i x¯ i qij where x¯ i is the international  average price, x¯ in is the average price p q of the benchmark country i x¯ijq ij is the value index with the same i i ij

5 See Cong Peihua’s “The Value Scale to Unite Volume and Value in International Economic Comparison—Analysis of the Defects of the UN ICP Method” Issue 5, 2007, p. 91.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

output structure and unified magnitude, and

 x¯ q i in ij i x¯ i qij

67

is the relative

price index obtained by weighting the output structure of the jth country. The GK method simultaneously eliminates the differences in the currency value and averages the relative price to obtain the conversion indices of the PPPs of different countries. As the PPP that is estimated by the GK methodincludes a conversion factor x¯ q that differs across countries, namely i x¯inq ij the PPPs computed by i i ij the GK method differ across the sample countries. The PPP comparison measures the real living standard of the ­residents of a country. From this perspective, the PPP can be regarded as the comparison of the total commodities formed by the consumption baskets of different countries. The total commodities have varying compositions across countries, and they directly affect the estimation of the PPP on the relative price indices of different countries. For example, when comparing a less-developed agricultural country with a developed industrial country, the per capita income of the latter, which is converted by the exchange rate method, is four times greater than that of the agricultural country (4:1). Therefore, the residents in the agricultural country spend 100% of their income on buying basic survival items (e.g., food), whereas the residents in the industrial country spend 20% of their income on food and other basic survival items. They spend the other 80% on luxury or capital goods.6 The average price of luxury or capital goods is five times that of food. In this case, the price of food is set to p and that of luxury or capital goods is set to 5p. When the commodity difference is ignored, the ratio of the commodity price in the industrial country to that in the agricultural country is 4.2:1, that is, (5p × 80% + 1p × 20%)/ (1p × 100%). As the per capita income of the industrial country is four times greater than that of the agricultural country and that the price level of the industrial country is 4.2 times greater than that of the agricultural country, the real living standard of residents in the industrial country is lower than that of the residents in the agricultural country after the price factor converted by the PPP is eliminated. The residents in the industrial country are 5% poorer than those in the agricultural country. This conclusion is clearly illogical. Even if these commodities are the same, 6 The assumption of this example is that the residents of both countries do not have savings. That is, they spend all their income on consumption.

68  X. SONG

their quality may differ across countries. For example, German made Mercedes Benz greatly differs from China’s Xiali in terms of quality. These arguments all reflect the limitations of the PPP method. 4.3.2   Three Methods for Calculating the PPP Similar to the accounting of the GDP, the PPP can be calculated using three basic methods, namely, the  expenditure method, production method, and income method. The values obtained using the production and expenditure methods can be divided into meaningful price and quantity components, and those that are obtained using the income method cannot be divided. Therefore, when choosing a method to calculate the price index of the PPP, only the expenditure and production methods are considered. The ICP mainly uses the expenditure method for calculating the PPP. This method not only enables the comparison between the main elements of the final demand (i.e., consumption and investment) but also avoids the difficulty in the production method. That is, the difficulty to obtain both intermediate consumption data and total output data in order to realize double deflation. The PPP calculated using the expenditure method is both a spatial price deflator that measures the relative difference in the product price between economies and a currency convertor that converts the GDPs expressed in different currencies to GDPs with a similar economic basis. To evaluate the expenditure method, we need to understand first the PPP calculation process using the expenditure method. In terms of personal consumption, PPP is divided into three levels in the geographic dimension (i.e., economy, great region, and globe) and five levels in the product dimension (i.e., basic heading, class, group, category, and main aggregate). The PPP of the basic heading of economy in great regions is obtained by calculating the ratio of the national average price of the product specifications in the regional catalog under the basic heading to the national average price of the basic heading of the reference country in the great region. The PPP of the basic class of economy at the global level is obtained by linking the interregional PPP calculated according to the price of the basic heading of the great region after deflating the prices of product specifications in the global catalog using the PPP of the basic heading of economy in the great region. The PPP of economy in the great region in the other product dimensions is obtained

4  VIEWING AND APPLYING THE PPP CORRECTLY 

69

by summing up the PPPs of the basic headings in the great region level grade by grade. The proportion of the expenditure of the corresponding product level obtained by decomposing the GDP of the economy can be used as the weight. The PPP of the economy at the global level in other product dimension levels is obtained by summing up the PPPs of the basic headings at the global level using the CAR method. The weight that is used in the summary maintains the fixed real expenditure proportion expressed in the global unified currency (i.e., the ratio of the expenditure of the basic heading expressed in the local currency to the PPP of the basic heading of economy at the global level). Unlike the products that consumers buy in the market, housing, health care, education, government compensation, building, and machinery and equipment are difficult to compare. Moreover, their data cannot be completely obtained through a market survey. Take, for example, the housing PPP calculation method for the ICP 2011. Computation methods varied across different regions. Africa, Latin America, the Caribbean, and West Asia adopted the economy product virtual method, the Asia–Pacific region used the benchmark method, the EU directly adopted rent data, other economies indirectly adjusted the quality and number of buildings, and the Commonwealth of Independent States used the quantity method. Countries in Africa, Latin America, the Caribbean, and West Asia used rent data to link housing PPP. The Asia–Pacific Region and the EU– OECD countries used dwelling stock data. Due to the different housing quality in the Asia–Pacific and the EU-OECD regions, this may have overestimated the housing PPP of the EU–OECD or underestimate the housing PPP of the Asia–Pacific region. Government compensation adopted the input cost method, but this method only investigates the compensation of employees in the production cost. Africa, the Asia–Pacific region, Latin America, and the Caribbean adjusted government productivity after estimating capital–labor according to education levels and the availability and utilization rate of the equipment. The EU–OECD, the Commonwealth of Independent States, and West Asia did not follow such a procedure because the adjustment coefficient of government productivity was greater than 1, which overestimates the output value of the adjusted regional government service. For construction, the EU used the bill of quantities, the Com­ monwealth of Independent States combined the input and output methods, and other regions used the weighted average input price of labor,

70  X. SONG

material, and equipment leasing. The input method assumes that the total factor productivity is consistent between economies. To link the construction PPP, Russia used both the bill of quantities and the hybrid method. Several EU–OECD countries used both the bill of quantities and the input method. The potential assumption of the input method underestimates the PPP of countries with high total factor productivity. The linking potentially assumes that the cost profit margin of construction in the EU–OECD countries is the same as that of construction in other regions. Given the high production cost of the construction industry in the EU, this linking underestimates the cost profit margin of other regions and undervalues the construction PPP in other regions. Therefore, when estimating PPP, several problems are observed in nonhousehold consumption. Clearly, all classifications tend to be overestimated the PPP in the EU and underestimate that in the Asia–Pacific region. 4.3.3   Defects of the Expenditure Method Based on the unity between household consumption and non-household consumption, the PPP calculated by the expenditure method has inherent defects. The first defect lies in the regional representativeness of the products in an economy. The production of a product is completed by enterprises. Given the enterprise collectivization, multiregional investments, and multinational businesses of today, the GDP statistics of an economy cannot accurately define national boundaries and may lead to repetitive GDP calculation and either overestimating or underestimating the GDP of a region. Therefore, statistical errors can exist in the GDPs of different class levels. The second defect lies in the product level representativeness of the products in a region. GDP represents the output of the economy at the product level instead of total regional consumption. If the unit consumption price is weighted with the total output, computing the prices of different classes by summing up the unit consumption prices will yield inaccurate results. The third defect lies in the economic representativeness of the product specifications in a region. The regional and global catalogs consider the comparability of the selected specifications in the region and in the world, respectively, but they do not consider the representativeness of the specifications in an economy. As a result, for regions with greater

4  VIEWING AND APPLYING THE PPP CORRECTLY 

71

differences in their resources and economic development, the product value in the regional catalog accounts for only a small proportion in the PPP, which consequently fails to represent the whole region. The fourth defect lies in the representativeness of the economies within a region. For example, the GDPs and per capita GDPs of countries within the Asia–Pacific region greatly vary. Therefore, the representativeness of an economy in this region is significantly weakened. In this case, the PPP of the Asia–Pacific region used as the basis of the global linking becomes unreliable. South Korea, Japan, Singapore, and Taiwan are some of the economies in the Asia–Pacific region with a high per capita GDP. South Korea and Japan are classified under the EU–OECD, while Singapore and Taiwan remain within the Asia–Pacific region. This classification implies the lack of a unified region division standard that leads to errors in estimating the PPP of the Asia–Pacific region. 4.3.4   Limitations of the Production Method The production method performs an international comparison from the production perspective and is based on the decomposition of the GDP by industry. ICP has systematically developed a methodology for internationally comparing price, output, and productivity from the production perspective and has made this method adaptable for whole national economies. This method is indirectly and slightly affected by price. Calculating with cost price can avoid price distortions caused by subsidies and other additional fees. The PPP calculated using the production method is actually a unit value ratio. The production method calculates PPP from two aspects. First, the production method calculates the PPP using the factory output value and the quantity of each product, both of which require the participating countries to collect as much product or service data as possible. Second, the production method calculates the PPP using the output value of each industry, which requires the participating countries to provide as many output values of industries as possible. The production method matches the comparable products or industries of two countries and then calculates the unit price ratio of each pair of matched products or industries. The matched products or industries must follow the comparability principle and be representative of their corresponding industries. The production method calculates the PPP as follows:

72  X. SONG

First, the value of a product is calculated according to its production value and quantity. The unit price of each product is its average price, that is, the average ex-factory price of the products of the same kind or of the similar products produced by several manufacturers in a country or region within a year. Second, the production method matches the products with the same characteristics between two countries and regions and then calculates the unit price ratio of each pair of matched products between two countries and regions. Third, the production method obtains the weighted summary of the unit price ratios by using the output quantities of the countries in comparison and of the benchmark country as the weights. The output quantities of the countries in comparison and of the benchmark country are weighted to obtain the PPP of the sample industry (a sub-class of another industry). Fourth, the PPPs of the sample industries are weighted to obtain the PPPs of different industries. The weights represent the total added value of the sample industry. Fifth, the weighted aggregates of the PPPs of all industries are used to represent the PPP of the whole national economy with the added value of the industry as the weight. In this way, the GDPs of countries in comparison are converted to those expressed in the currency of the benchmark country. The weakness of the production method lies partially in its way of calculating the PPP of the service industry. The method thus has been modified to account for the unique characteristics of the service industry. The output accounting of the service industry is divided into two parts, namely the market and the non-market services, which adopt the cyclic and input methods, respectively. Similar to Step 3, as the prices of most services cannot be directly obtained from census data, the weighted aggregates of the quantity indices are used to obtain the output PPP of the sample industry by using the output values of the sample industries from the countries in comparison and from the benchmark country as the weights. The price multiplied by the quantity represents the output. As the PPP of the service industry is calculated on the basis of its output quantity index, the output quantity is the basis for calculating the output PPP. However, calculating the output quantity of the service industry is difficult. The output index of some non-market services has to be replaced with the output index in the

4  VIEWING AND APPLYING THE PPP CORRECTLY 

73

service process. For instance, the medical service output is replaced with the number of patients who avail themselves of medical services. As some outputs are combinations of several output indices, these indices must be classified and integrated through quality adjustment.

4.4  Basic Application of ICP 4.4.1   International Economic and Structural Comparison The PPP data of the ICP have been widely applied by international organizations in their analysis and research. This project also provides international organizations and countries with globally comparable economic statistics and analysis techniques that can help them study world economic development, monitor the achievement of millennium development goals, and make relevant decisions. For example, it is used for calculating the global poverty rate (World Bank) and the global economic growth rate (IMF), allocating structural funds (EU), evaluating human development index and gender equality (UNDP), health inequality (WHO), per capita education expenditure (UNESCO), monitoring the living condition of children (UNICEF), and aiding the project designs of other international organizations. The data from this project have also been widely used in investigating world economic competitiveness, investment cost, potential industrial growth, and adjustment of living expenses in different cities. PPP is basically applied in international economic and structural comparison. The ICP produces internationally comparable GDP data and the price and quantity measurement results of its components. As these results are calculated with the PPP as the currency conversion factor, PPP is at the core of the ICP. The ICP aims to calculate the PPPs of different countries and then use the results as the currency conversion factor for converting the GDPs, consumption, capital formation, and net exports expressed in the local currency to those expressed in a unified currency to compare and evaluate the actual economic scales and structures of different countries. Take China as an example. According to the ICP 2005 results published by the World Bank in December 2007, the PPP of China was 3.40 (CNY/USD), which indicated that the PPP of $1 was equivalent to that of 3.40 CNY. In 2007, the GDP of China amounted to 26.58 trillion CNY, and the exchange rate of the US dollar against the RMB was

74  X. SONG

1:7.62. According to the exchange rate method, the GDP of China was $3.49 trillion after conversion. However, according to the PPP method, the GDP of China was $7.82 trillion after conversion, which was 2.24 times more than the former. Clearly, the economic scale calculated using the conversion factor of the PPP is greater than that calculated using the exchange rate method. In the recent ICP 2011 results published by the World Bank, the PPP conversion factor of China was 3.51 in 2011 (CNY/USD). According to the National Bureau of Statistics of China, the GDP of the country in the same year was 47.31 trillion yuan, which amounted to $13.50 trillion after conversion. In the same year, the average exchange rate of the US dollar against the RMB was 1:6.45. After conversion according to the exchange rate method, the GDP of China was $7.33 trillion, which was 1.84 times less than that calculated according to PPP. 4.4.2   Price Level Comparison The price level index (PLI) is computed by dividing the PPP of a country by its exchange rate. This index is often used to reflect the difference between the domestic price level of a country or a region and the international price level (Yu 2008). PLI has the following applications: First, the PLI is used as the standard to decide whether dumping exists in international trade. According to the regulation of the World Trade Organization, if the export price of a product is lower than the comparable price of the same product consumed domestically, then dumping has occurred. Second, the PLI is applied to determine and adjust the standard of living abroad. ICP and PLI data are widely used to determine the pay and living standards as well as to adjust the subsidies and allowances of the expatriate employees of multinational companies, NGOs, and the Agency for International Development. These data are also used by the International Labor Organization (ILO) to compare the minimum wages and adjust the living costs in different cities. Third, the PLI is used to evaluate the investment costs across countries. Multinational companies are increasingly using ICP data to monitor their overseas investment costs to determine the viability of their projects, evaluate the labor and material costs in their international operations, analyze their market share, and seek direct investment opportunities.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

75

Fourth, the leaders of multinational companies use the PLI to analyze the comparative advantages of goods and services in terms of price and expenditure. Policy makers also use these data to evaluate their competitive advantages in international trade. Specifically, comparing the price levels of GDP components between trade partners helps policy makers determine whether their price level is reasonable and conforms to the requirements of their trade partners to prevent international trade imbalance. Fifth, PLI is used to determine compensation standards. When designing transnational compensation standards, compensation management organizations are always concerned whether their products are reasonably priced compared with those of the target countries. The PPP and PLI indices of the basic product classes can provide the judgment standard. In addition, PPP can also be used to compare the price levels among regions, namely the regional price difference index. The statistics departments of the United States, the UK, Canada, Australia, and other countries have tried to compile the price difference indices among regions. Some international organizations and experts measure the price difference degrees among regions and between the rural and urban areas of large countries. The US Bureau of Economic Analysis and the US Bureau of Labor Statistics update their calculation results every year and regularly publish research reports in Survey of Current Business. In 2003, the UK expanded this survey to the regional Relative Consumer Price Levels project (RRCPL), which calculates the price difference indices of five regions, including London. The survey results for 2004 and 2010 are publicly available. The Canada Dominion Bureau of Statistics calculates and officially publishes the retail price difference indices among 11 large cities in the country (the national capital and 10 provincial capitals) and compares the living costs among cities every year. The comparison results are used as the basis for determining and adjusting the low-income standard and poverty line among regions. When studying poverty and income distribution problems, international organizations, experts, and scholars focus on the price difference among regions and between urban and rural areas. However, the statistical agencies of most developing countries have neither systematically performed regional price difference index research nor officially released relevant official data. Based on the ICP results, the World Bank estimated that the price difference among regions in some developing

76  X. SONG

countries was less than 5%. The 2005 ICP survey data revealed that when the quality and income effect factors of the unit value of food were adjusted, the price difference between rural and urban areas in India would only be 3.2%. In their paper Spatial Price Difference within Large Countries, Angus Deaton et al. from Princeton University calculated the price difference degree of food among the main regions and between the urban and rural areas in India and Brazil (Deaton and Dupriez 2011). In 2010, the World Bank and the ADB employed the CPD method to calculate the price difference indices of 17 regions in the Philippines based on the price and household expenditure data from the country’s CPI survey. Chinese scholars have recently investigated the measurement of price differences among regions. In 2005, Jiang Xiaojuan and Li Hui selected the prices of 22 specifications in 7 classes of expenditures from 36 Chinese cities to measure the relative price of each commodity in Beijing. The average of these relative prices was then taken as the total price level among cities. The price levels of 36 cities showed obvious differences that were closely related to the per capita income level of these cities. By applying the regional price difference index method, a Chinese research group (2014) calculated the price levels of first-level regions using the Jevons index, summed up these levels using the Paasche index, and measured the differences in the per capita income of regions using the Gini coefficient. The income differences among regions demonstrated a narrowing tendency between 1995 and 2004. Wang Lei and Zhou Jing (2012) constructed the general spatial CPD model to calculate the relative price level indices by using the consumption expenditures of residents in 31 provinces (regions and cities) of China. They found that the spatial CPD model was highly suitable for estimating the relative price level between Chinese regions and that an inverted U-shaped curvilinear relationship was found between the price level difference and the market integration degree. 4.4.3   Comparison of Productivity Among Departments After calculating the departmental PPP conversion factors, the PPP method is also used to compare productivity across departments. Paige and Bombach (1959) compared the departmental productivities of the UK and the United States, and this comparison marked a significant advancement in the use of the production method for international comparison and laid a foundation for subsequent research. In the mid-1990s,

4  VIEWING AND APPLYING THE PPP CORRECTLY 

77

some scholars compared the productivity levels of individual industries, such as manufacturing, agriculture, mining, transportation, and business, but only few of these studies considered the entire national economy as the comparison object. Global ICP PPP is calculated based on expenditure. Using the production method to calculate PPP requires that the weight must be selected from the producers during calculation. In 1983, the University of Groningen in the Netherlands established the international output and productivity comparison project and began performing systematic research on the production method. The project systematically developed a methodology for the international comparison of price, output, and labor productivity from the producers. Under the ICP framework, Pilat (1994) compared the departmental outputs and productivity levels of Japan and South Korea by using the whole national economy as the object. 4.4.4   Industrialization and Economic Development Level Analysis The ICP calculation results can also be used to analyze economic development. Gilboy and Zhong Ninghua (2010) constructed a relative price index of machinery and equipment (the ratio of machinery and equipment PPP to real individual consumption PPP) to reflect the industrialization level of a country. They contended that the PLI (the ratio of the PPP to the exchange rate) reflected the difference between domestic and international price levels and had certain guiding significance. By using the ICP 2005 results published by the World Bank, Luo Zuchun and Gao Bo (2009) proposed new standard values of per capita GDP in different stages of industrialization and analyzed the standard recognition in different stages of industrialization 60 years after the founding of the People’s Republic of China. The change in the main statistical system provided empirical evidence for evaluating the industrial growth potential and scientifically proposing urbanization policies. The PPP data are also used in investigating economic development and structural change. For example, the International Monetary Fund (IMF) uses the PPP to estimate the world and regional economic growth potential. Given the close relationship between economic growth and investment, the proportion of GDP investment can be regarded as a core index for measuring the economic growth potential.

78  X. SONG

4.4.5   Poverty PPP and Poverty Statistics Poverty standard research presents an important direction for the application of the ICP method and data. During the Millennium Summit in 2000, the UN formulated a millennium development goal that prioritized the elimination of extreme poverty and hunger. In 2011, the eighth round of the ICP included three goals, one of which was to calculate poverty PPP data that could be used as the statistical bases for monitoring poverty alleviation. Poverty PPP refers to the ratio between the purchasing powers (in two or more currencies) of a certain number of typical products that are consumed by poverty-stricken people. Unlike the PPP that targets all residents in a country, poverty PPP focuses only on the poor people of a country. The poverty line that is based on consumption or income is the most widely accepted poverty standard in the world. The poverty line indicates the minimum income or consumption level required for an individual to sustain life. Based on the poverty status of 33 developing countries, the World Bank marked $1/day as the extreme poverty line, $2/day as the poverty line, and then estimated the global poverty rate (World Bank 2013). Based on the PPP index in 2008, the ADB proposed $1.35/day as the poverty line for Asia. Following the advice of World Bank experts, the poverty line of China is calculated according to the minimum consumption required by a person per day (adjusted every year). Some scholars calculate the rural poverty standard and status of China according to the 2005 PPP data and found that the rural poverty standard at the time was close to $1/day. In the same year, the rural poverty line of China was a per capita annual net income of 683 CNY. Since then, the poverty line of China has been raised continuously from 693 CNY in 2006 to 1196 CNY in 2009 and 2300 CNY in 2011. As a problem that closely accompanies the study of the poverty standard, poverty measurement also requires the use of PPP data. The World Bank and the US Carnegie Endowment for International Peace previously used the PPP to conduct research on poverty measurement. Measuring inequality is another important application of the PPP data. For example, WHO and UNESCO use the PPP to measure the inequalities in per capita expenditures on health and education, respectively. UNEDA uses the gender power measure index to measure gender inequality between two countries.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

79

4.4.6   Comparison of International Competitiveness and Living Standards The price decision mechanism of many products within some economies is isolated from the international market. When evaluating the living standard, international competitiveness, or productivity of an economy, the PPP overcomes the flaws of the exchange rate method (i.e., distortion and frequent changes in conversion) that are caused by the balance of payments, financial market, export strategy, and mentality. By reflecting the differences in the price level of countries, the conversion rate that is provided by PPP can help people compare the actual total output between countries more accurately and then compare the living standard between countries according to the actual per capita consumption calculated based on the PPP. The GDP converted by PPP is often used to measure international economic competitiveness. Ren Ruoen et al. (2006, 2008) examined international competitiveness by using the PPP result and calculation method. 4.4.7   Application in Other Aspects The budget and evaluation of R&D funds must consider the price factor. If the prices of a country are lower than those of other countries, the same amount of money has greater purchasing power in countries with low prices. Although using the PPP method to examine the purchasing power of R&D funds from the price perspective is different from using the absolute level measure of the funds input, the results may have higher practical significance. Clearly, the PPP data in the ICP results can reasonably measure the R&D funds scale and provide reference for investment decisions. The household consumption expenditure in the ICP results is combined with the CPI data. Measuring the contributions of the household expenditure components to CPI change can provide policy suggestions on stabilizing prices and controlling inflation. The EU has recently developed a policy for distributing EU structural funds by using the PPP-calculated per capita GDP. USAID also uses the PPP-calculated GDP to evaluate the eligibility for aid. Given the increasing maturity of the ICP, the PPP may be used to calculate the membership fees of member countries, evaluate the eligibility for aid and donation, or determine the preferential terms for obtaining loans.

80  X. SONG

4.5  ICP in China 4.5.1   Progressive Development and Gradual Expansion Although China did not formally participate in the ICP before 2005, the country actively cooperated in the project. The National Bureau of Statistics of China has established an International Statistical Information Center that is responsible for ICP information exchange and cooperation. Many Chinese scholars translated, introduced, and conducted in-depth investigations of ICP literature. For example, Prof. Ren Ruoen from the Beijing University of Aeronautics and Astronautics and Prof. Guo Xibao from Wuhan University calculated the PPP of RMB against the US dollar based on the GDPs and per capita GDPs of China in 1986 and 1994, respectively. In 2005, about 11 Chinese cities (i.e., Beijing, Shanghai, Chongqing, Dalian, Ningbo, Xiamen, Qingdao, Harbin, Wuhan, Guangzhou, and Xi’an) participated in the Asia–Pacific ICP survey of the World Bank (2005 was the benchmark year). The accounting scope of GDP expenditure was expanded from parts of cities to the whole country, and the price survey scope was expanded to rural areas. The World Bank calculated the national average price level and the basic classes of GDP expenditure by using the price data of these 11 cities and other relevant statistics. The calculation results were then used as the basic data for calculating the PPP of China. In 2011, China participated in the eighth round of the global ICP and performed a price survey in 30 provinces (regions and cities) to calculate the GDP expenditure by basic headings. 4.5.2   PPP Conversion Factor and Price Ratio of China The World Bank database has published the PPP data of China since 1990. As the calculation method of the ICP refers to the PPP of a non-benchmark year, each round of ICP has an internally consistent PPP every year, but the PPPs between two adjacent rounds are not consistent. According to a certain model, the PPP in the benchmark year of the latest round is taken as reference to adjust further the PPP of each round and to achieve consistency in the PPPs between two adjacent rounds. Table 4.4 shows the PPPs of China over the years with 2011 as the adjustment reference. From 1990 to 2013, the PPP conversion

4  VIEWING AND APPLYING THE PPP CORRECTLY 

81

Table 4.4  PPP conversion factor of China, GDP Year

PPP (LUCa/international dollar)

Year

PPP

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

1.63 1.78 1.97 2.21 2.59 2.83 2.93 2.90 2.81 2.73 2.74 2.74

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

2.71 2.73 2.84 2.86 2.88 3.02 3.19 3.15 3.32 3.51 3.51 3.52

aLUC means Local Unit Currency Data Source World Bank database

Table 4.5  International rankings of China in terms of PPP and the relative price level

2005 2006 2007 2008 2009 2010 2011 2012 2013

PPP

Ranking

PPP/EX

Ranking

2.86 2.88 3.02 3.19 3.15 3.32 3.51 3.51 3.52

83 80 81 79 79 78 78 77 76

0.35 0.36 0.40 0.46 0.46 0.49 0.54 0.56 0.57

67 59 61 73 84 91 103 111 116

Note Rankings are based on data from the World Bank database for all 172 economies after excluding the economies with missing data

factor of China increased twofold from 1.63 to 3.52. The PPPs of China increased quickly between 2006 and 2008 and between 2010 and 2011 but decreased slightly in 2002 and 2009. The data on relative price level show that the international ranking of China in terms of relative price level (PPP/EX) dropped from 67th in 2005 to 116th in 2013. In fact, the price level in China has been steadily

82  X. SONG

increasing during this period. The reason for the drop in ranking is RMB appreciation. In other words, $1 can be converted into a lesser amount of the RMB, so that the increase in the relative price level greatly exceeded the change in the PPP. If the currency conversion ratio is expressed by the exchange rate, the relative price level can be expressed by the reciprocal of PLI. Table 4.5 shows the international rankings of China in terms of the PPP and the relative price level. 4.5.3   PPP-Converted GDP of China The GDP of China can be easily converted from LUC to US dollar by using the PPP conversion factor (see Table 4.6). The ICP 2005 results for China showed that the PPP of the country was 3.45 in 2005, which means that $1 was equal to 3.45 CNY. The PPP was equivalent to 42.1% Table 4.6 Chinese GDP in LUC and in US dollar calculated by the PPP conversion factor

Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

GDP (LUC) (100 million CNY)

GDP (PPP) (100 million USD)

18,667.82 21,781.50 26,923.48 35,333.92 48,197.86 60,793.73 71,176.59 78,973.03 84,402.28 89,677.05 99,214.55 109,655.17 120,332.69 135,822.76 159,878.34 184,937.37 216,314.43 265,810.31 314,045.43 340,902.81 401,512.8 473,104.05 519,470.10 568,845.21

Data Source National Bureau of Statistics of China

11,426.67 12,260.00 13,690.58 16,012.42 18,636.92 21,514.37 24,300.67 27,190.20 30,040.45 32,856.24 36,163.28 40,062.27 44,372.28 49,796.40 56,321.08 64,701.76 75,144.86 88,063.86 98,434.58 108,331.98 121,097.73 134,959.12 147,826.97 161,577.04

4  VIEWING AND APPLYING THE PPP CORRECTLY 

83

of the exchange rate in that year. According to the PPP method, the GDP of China was $5.3332 trillion, which accounted for 9.7% of the global GDP. By contrast, by using the exchange rate method, the GDP of China was $2.2438 trillion, which accounted for 5.1% of the global GDP. The PPP of China was 3.51 in 2011, which means that $1 had the same purchasing power as 3.51 CNY during the year. This PPP was equivalent to 54.3% of the USD/CNY exchange rate (6.46) in the same period. China was the second-largest economy in the world in 2011 according to the GDPs calculated by both the PPP and the exchange rate methods (accounting for 14.9% and 10.4 of the global GDP, respectively). Compared with the ICP 2005 results, the 2011 ranking of China in terms of GDP computed by the PPP method remained unchanged, but rose from fourth to second in terms of GDP computed by the exchange rate method. According to the international ranking of PPPs, between 2005 and 2013, China rose from 83rd to 76th, Brazil dropped from 52nd to 57th, India from 105th to 108th, and Russia from 111th to 116th. PPP rankings of the latter three countries all showed a downtrend.

4.6  Challenges Faced by the ICP 4.6.1   Flaws of the PPP Method in Theory Logically, theories might be perfect, but reality is often different from perfection. Each theory and statistical method has its advantages and disadvantages. Some commonly used concepts in economics, such as potential growth rate and full employment, seem to be perfect and highly significant in theory, but in reality, these concepts are unpredictable. In other words, a quantitative economist or statistician cannot easily obtain the actual values of potential growth rate and full employment. Prof. Eugene Fama from the University of Chicago won the 2013 Nobel Prize in Economics for his famous efficient market hypothesis (EMH) in the capital market. Although this hypothesis is logically perfect, the situation in the real market is far from the EMH. A perfectly efficient market is unattainable.

84  X. SONG

The GDP converted using the exchange rate method is widely known to be misleading in showing the relative size and material wealth level of an economy. Moreover, the exchange rate method does not distinguish trading from non-trading commodities when converting the GDP expressed in the local currency to that expressed in the unified currency. Both of these commodities only follow one exchange rate. Processing the data of non-trading commodities or services using a single exchange rate leads to the assumption that these commodities or services are tradable. However, such assumption lacks a realistic foundation. The GDP converted using the PPP method does not show this deviation because the PPP method initially calculates the price ratios of individual products. Therefore, the differences in the price levels of trading and non-trading commodities are considered in the calculation. Although the PPP theory is logically perfect, several challenges in its method and practice must be addressed to calculate the true PPP of a country or region. The PPP method cannot truly reflect the price level ratio because the price level is affected not only by the quantity level but also by quality, consumer habits, and geography. Quantity is also influenced by quality. If the eliminated price level contains the quality factor, the remaining quantity level becomes incomparable. To perform the comparison, homogeneity must be emphasized and the quality factor must be eliminated. However, the PPP method cannot adjust the quality of all commodities and services. To avoid this limitation, the product specifications of the same or similar quality are selected for comparison. However, the importance of the specifications as well as their representativeness in the basic classes may vary across each country. In this case, the representativeness of the specifications should be addressed immediately. The PPP method only considers the importance of basic classes in ­different economies and then distinguishes the important specifications from the unimportant ones. However, by ignoring the representativeness of these specifications in their respective economies, the PPP cannot fully reflect the price level ratio without the quality factor. The theory of ICP seems perfect, but it still contains some inherent flaws (Yuan et al. 2008). First, Gustav Cassel believes that the exchange between two currencies depends on the currency purchasing power. The PPP theory is based on the traditional quantity theory of money but largely ignores the system, trading cost, technology, and other factors. Second, ICP does not distinguish trading commodities from non-trading ones and ignores non-trading factors, such as trade costs and barriers.

4  VIEWING AND APPLYING THE PPP CORRECTLY 

85

Third, the ICP emphasizes the effect of prices on the exchange rate. However, changes in the exchange rate may affect prices in turn. The disagreement between the exchange rate and the PPP methods primarily lies in their different potential assumptions. If all commodities and services are circulated all over the world, the currency supply and demand is driven by international trade, and the exchange rate can fully reflect the currency ratio, then the price level ratio will automatically eliminate the influence of consumer habits and spatial geography. According to the law of one price, the exchange rate and PPP methods would be equivalent. We must rationally look at the ICP and the PPP. Perhaps such imperfection may represent the beauty of economics and even of human society. 4.6.2   Needed Improvements in the Statistical Method of the ICP Given the problems of the ICP in its method and practice, especially its overestimation of the GDPs of developing countries, many developing countries have become skeptical of the project. Some countries joined the project but eventually discontinued their participation. Therefore, the measurement of the PPP must be analyzed rationally. The ICP results have not been actually used in making administrative decisions because the project still has flaws in its methodology. As its greatest challenge, the project must determine how to design a relatively scientific statistical method that considers different country sizes, cultures, commodities and services, and statistical capacities and that can generate accurate and reliable PPP (Yuan et al. 2008). Affected by the policy systems and economic, social, and cultural conditions of different countries, the ICP still has many technical problems that have not been effectively solved during its implementation. These problems include the designing of an internationally comparative and representative price survey framework; performing international comparison in construction, machinery and equipment, housing, public education, and healthcare; aggregating and linking the global PPP; calculating the poverty PPP; calculating and fixing the PPP in the non-benchmark year; and determining the representativeness and comparability of specifications. These problems have always been present in the implementation of the ICP and have affected the data quality of the PPP. The ICP has many shortcomings in aggregating data, setting product specifications, and conducting

86  X. SONG

price surveys. All these require further improvements to ensure the reliability of its results. First, PPP data aggregation methods can lead to deviations. The current ICP has many investigation, aggregation, and calculation methods that can lead to different results. In practice, no reliable empirical evidence can support one method over another. For example, the PPP of a single commodity or service can be easily computed by directly comparing the prices of relevant commodities among countries, since the PPPs of single commodities have a natural transferrable property among countries. However, the statistical process of classifying single commodities into basic heading and higher-level classes has many contradictions. Calculating the PPPs of basic headings according to the PPPs of single commodities or services is actually calculating the geometric average of the PPPs of all commodities or services in each basic class. Other PPP calculation methods include the EKS method, the CPD method, and the GK method. The characteristics of these methods are shown in Table 4.7. The comparison results are sensitive to the calculation method. Therefore, using different methods can lead to different results. The rankings, absolute values, and relative gaps in the GDPs per capita Table 4.7  Advantages and disadvantages of different PPP methods Method

Applicable situations

Advantages

Main disadvantage

EKS

PPPs for basic headings or aggregation of basic heading PPPs

Only use part of the data for countries in comparison Does not meet matrix consistency requirements

PD

PPPs for basic headings

GK

Aggregation of basic heading PPPs

Convert intransitive binary index to transitive and multilateral index Can fill in missing values in binary PPP matrices Can better present expenditure structure of all countries Especially suitable for cases where price data are partially missing Satisfy transitivity Not affected by the choice of the reference country The international prices have clear economic significance Make full use of data Guaranteed additivity Guaranteed transitivity

Subject to the Gerschenkron effect The missing data approaches have no economic significance Subject to the Gerschenkron effect Results are affected by the adjustment of the underlying data

4  VIEWING AND APPLYING THE PPP CORRECTLY 

87

between countries and between the national and the regional level may differ when various methods are adopted. Therefore, further research must be conducted to determine a method that can accurately reflect the economic strength of countries. To develop and improve the theory and methods of ICP, the UN and the World Bank formed the ­technical advisory group (TAG) in the global ICP Commission. ICP technologists from all over the world were invited to solve problems in ICP techniques. The comparison results are sensitive to the calculation method. Therefore, using different methods can lead to different results. The rankings, absolute values, and relative gaps in the GDPs per capita between countries and between the national and the regional level may differ when various methods are adopted. Therefore, further research must be conducted to determine a method that can accurately reflect the economic strength of countries. To develop and improve the theory and methods of ICP, the UN and the World Bank formed the TAG in the global ICP Commission. ICP technologists from all over the world were invited to solve the problems in ICP techniques. Second, when selecting the representative specifications, the contradictions that result from the differences in consumption level, consumption structure, and commodity quality among countries cannot be coordinated easily. In particular, the consistency in and comparability of these specifications cannot be easily realized. For example, haircut services in China and the United States cannot be easily considered as homogenous. Given their differences in traditions, habits, economic development level, and economic structure, the consumption structures of various countries also greatly differ. The ICP selects commodities by mainly using Western developed countries, particularly the United States, as the standard. These specifications meet the principle of representativeness in Western developed countries but not in developing countries, and thus their comparability is reduced. Third, the ICP is based on national economic accounting and requires detailed composition data of GDP expenditure classifications that cannot be easily obtained from countries with lower statistical capacity. Some countries may deliberately distort their basic data on prices and GDP expenditure classifications to secure their political and economic interests, and thus affecting the reliability of the ICP comparison results. Fourth, the quality of commodities may cause problems. The differences in the quality of commodities are reflected in two aspects, namely

88  X. SONG

the difference in physical properties such as commodity specification and function, and the difference in intangible aspects such as marketing environment and service quality. This difference directly affects the price level of commodities. Overall, developing countries have a poorer marketing environment and a lower service level (e.g., after-sales service) than developed countries, and their prices contain a low added value of service. Excluding this factor may distort the prices of commodities in developing countries. Fifth, the issue of regional price difference can be easily addressed in smaller countries. However, for large countries such as China, the market prices of the same commodity may greatly across different regions. In addition, state-controlled planned price and price distortions still exist in some fields and may result in restricting the application of the PPP method in China. Sixth, data from the service sector show huge differences. As tax is the main source of statistical data, imperfect taxation systems in many developing countries leads to huge losses in service sector data. Some service commodities in many countries are not provided according to the market prices but are subsidized by the government, such as health care, education, and housing. The varying subsidy systems of different countries may lead to service price distortions. As the ICP cannot effectively deal with some of these price distortions, the project adopts the input method in which the salaries of teachers and doctors are compared as representatives to eliminate or reduce the distortions. As a result, the service price level of developing countries can be easily underestimated because of the huge gap in the capital formation and labor productivity between developed and developing countries.

4.7  Misconceptions About the PPP People hold various misconceptions about the PPP method and the ICP. The most common misunderstandings are discussed below. 4.7.1   Misconception 1: PPP and the Nominal Exchange Rate One of the most common misconceptions is taking the PPP value to determine whether the nominal exchange rate level is overestimated or underestimated. However, PPP can neither be regarded as the standard for evaluating whether the nominal exchange rate is overestimated or

4  VIEWING AND APPLYING THE PPP CORRECTLY 

89

underestimated nor be explained as the equilibrium exchange rate. Some people use PPP data to measure the deviation degree of the exchange rate level and the adjustment direction of the exchange rate. They simply consider that the currency is undervalued and needs to appreciate if the exchange rate is higher than the PPP. This view is based on an outdated theory of the PPP–exchange rate determination (Yuan et al. 2008), which is feeble in theory. Limited by the human knowledge level, the soundness of an economic theory is relative. In application, both time and space factors must be considered. Introduced in the 1920s, the PPP theory may be used to explain the exchange system in the gold-standard era and the Bretton Woods system. However, significant changes occurred in the international financial environment over the past century. After the collapse of the Bretton Woods system, the US dollar abandoned its peg to gold, and the world entered into a diversified age with a floating exchange rate. In this environment, only using PPP can not explain fluctuations in the exchange rate. Jacob Frenkel, former chief economist at the IMF and a professor at the University of Chicago, analyzed these flaws in using the PPP to explain the exchange rate problem. The exchange rate mainly reflects the parity of a country’s currency to other currencies in the mutual exchange of tradable commodities in the international market. The rate is decided by the actual social purchasing power level represented by the currency and the supply and demand relationship of the international market for commodities and currencies. Under specific conditions, the exchange rate reflects the price relationship of tradable commodities among countries. However, the calculation of the PPP involves many comparisons among non-tradable commodities. Because these commodities show great differences in terms of type and quality, non-tradable commodities are difficult to compare. In the data generation mechanism, the PPP covers all goods and services of the GDP, including tradable and non-tradable commodities, whereas the exchange rate only covers tradable commodities. These models also have different statistical scopes. Given such differences in the statistical scope and the influence of trading costs, the ideal PPP index may differ from the equilibrium exchange rate and cannot determine the real exchange rate. Theoretically, the PPP calculated by ICP is the product of a compromise of various aspects, and the ideal level of this PPP cannot be easily achieved. The ICP results published by the World Bank clearly show

90  X. SONG

that as a statistical estimation, the PPP index is affected by sampling error, measurement error, and classification errors and only serves as an estimate of the unknown true value. Therefore, countries must be cautious when using the PPP. Given the differences in methods and statistical principles, the PPPs from different rounds of ICP should not be compared against one another, and the economic indices converted by the PPPs cannot be used as the optimal comparison standards. In sum, the estimate of the PPP cannot be used as the criterion for evaluating the exchange rate. The exchange rate constantly fluctuates due to factors such as supply and demand relations and geopolitical factors in the international monetary market. The exchange rate fluctuates widely in some years but narrowly in other years. The global foreign exchange market is highly developed, and currency has become an important financial investment product. Transactions in the currency market are very active, and the daily turnover of foreign exchange can reach trillions of dollars. Factors such as the international economy, politics, cultures, and climate may all affect the nominal exchange rate level among currencies. The PPP value calculated by the ICP has no necessary correlation with the real exchange rate and their values may greatly vary. The changing direction and the magnitude of the PPP and exchange rate have no direct

Fig. 4.1  Trends of China’s PPP and the average exchange rate of the RMB 1990–2015 (Source Exchange rates from China Statistical Yearbook, and PPP data from http://databank.worldbank.org/data/views/variableselection/selectvariables.aspx?source=wo)

4  VIEWING AND APPLYING THE PPP CORRECTLY 

91

connections either. Therefore, the PPP cannot be regarded as the benchmark for evaluating the exchange rate level. As shown in Fig. 4.1, the trends of the PPP and the average exchange rate of the RMB against the US dollar are inconsistent. The exchange rate of the RMB against the US dollar continued to depreciate before 1994, was pegged to the US dollar in 1994 and 2005, appreciated in 2005 and 2008, was pegged to the US dollar again in 2008 and 2010, and appreciated after 2011. However, the PPP index of China continues to increase. In 2011, the exchange rate between the US dollar and the RMB was 6.46, and the PPP conversion factor was 3.51. As the two indices have different sources and problems, the exchange rate of RMB can neither be considered undervalued nor be taken as the excuse for requiring the appreciation of the RMB. 4.7.2   Misconception 2: Comparability and Representativeness of Comparison Objects The PPP method is used to calculate the GDP expenditures of different countries and to reflect their economic sizes according to the same price level. Therefore, the World Bank created a manual for implementing ICP 2011 and introducing the theory and methods of the ICP: Measuring the Real Size of the World Economy: Framework, Methods, and Results of the ICP (World Bank 2013). In theory, the PPP method outperforms the exchange rate method in terms of accuracy, stability, and ability to reflect the actual economic development level of different countries when performing international economic comparison. However, in reality, the statistical calculation of the PPP is a difficult process. Aside from the complex statistical method (i.e., the index structure, aggregation, and link methods have always been a major problem at the academic level of the ICP), many other challenges can hinder the implementation of this index. The greatest challenge lies in the choice of products participating in the comparison. On the one hand, these products must be comparable (i.e., can be found in all countries in comparison). On the other hand, these products must be representative of their home countries. Given the great differences in the society, economy, culture, politics, and geography among countries, either comparability or representativeness can be realized in practice. Products that are comparable among countries may not be representative in some countries

92  X. SONG

in comparison, and the representative products in a country may not be comparable with those in other countries. For example, the information technology industry accounts for a large proportion of the GDPs of Western countries but only accounts for a small proportion of the GDPs of many developing countries. In the rural areas of many developing countries, scattered individual households breed small numbers of livestock, but this mode of production has long disappeared in developed countries. As the production sets of two countries may intersect, the relative price between the same products of two countries can be naturally determined. However, many products have no overlap, and the portion of these products is greater among countries that are remarkably different in terms of their economic development (see Fig. 4.2). The relative price cannot be found if some commodities and services between two countries have no intersection. Therefore, countries have to divide their commodities or services into many groups and then select a commodity or service in each group as representative. If the commodities or services without intersection among countries only account for a small proportion, such approximation or ignorance can be accepted. Otherwise, the approximate statistics will result in a significant deviation. When collecting the ICP data, the specifications must be comparable and representative. However, comparability is often emphasized more than representativeness when selecting specifications. These specifications must be comparable regardless of their representativeness in their respective home countries. As China has a vast territory with great regional difference, selecting the specifications of different groups of commodities and services is a difficult task.

Fig. 4.2  Intersection of commodities and services between two countries

4  VIEWING AND APPLYING THE PPP CORRECTLY 

93

Many studies conducted by Chinese and foreign scholars show that the excessive pursuit of comparability can lead the ICP to overestimate the currency purchasing power, economic size, and per capita of developing countries. In fact, the ICP results can only be used to approximately categorize countries and regions instead of develop definite ranking among countries or regions. Moreover, the economic size converted by the PPP must not be taken as the standard for evaluating the economic strength of a country or a region. 4.7.3   Misconception 3: PPP and the Economic Structure Comparison When comparing two economies, some analysts also apply the PPP to compare various departments or economic structures. However, the PPP is inapplicable in this case. When applying the statistical method, the PPP can be used to compare the overall economic size of countries instead of their economic structure. The PPP is a comprehensive and overall index. Its calculation process involves the comparison of various commodities and services and covers all departments of an economy. When comparing the economic structures of countries, the value of different components may greatly vary from the real structure after conversion with the unified PPP value given the different composition proportions of tradable and non-tradable commodities. When analyzing the economic structures of different countries or regions (e.g., the relationship between the consumption and the government expenditures of different departments), the corresponding departmental PPP value must be used instead of the PPP of the overall economy to compare the departments of economies. 4.7.4   Misconception 4: PPP and the Poverty Measure The PPP is not a suitable measure for determining the poverty levels of different countries. Some studies have adopted the $1/day standard calculated using the PPP method (adjusted according to the price levels of different countries) to measure the poverty rate. This method is not rigorous and may greatly differ from reality for several reasons. First, the structures of commodities consumed by people living in poverty differ across countries. Unlike low-income people in highly market-oriented countries, the low-income population in some developing countries have a high degree of self-sufficiency. Second, theoretically, the PPP compares

94  X. SONG

the purchasing power of all commodities and services in the whole economy, but the structures of commodities and services consumed by people living in poverty are different from the structure of the whole country. Ignoring this difference will seriously distort the total value of commodities and services consumed by the low-income population. Therefore, different countries have various national conditions, and different regions and populations within a country have various income levels and consumption modes. In this case, the poverty rate must be measured according to the specific circumstances of countries instead of following a simple standard such as the PPP. Theoretically, residents with different income levels must be separated when calculating the currency conversion factor, which is used as a measure of poverty. This is a difficult task in terms of sampling, data collection, and data integration.

References Chinese References Gilboy, George J., and Ninghua Zhong, Measuring China’s Economy: The Proper Use of PPP Methods, Economic Research Journal, 2010(1). Luo, Zuchun, and Bo Gao, Determination of the Standards for Industrialization Development Period in New China in the Past 60 Years—Based on Study of the Results from the World Bank’s 2005 International Comparison Project (ICP), Academia Bimestris, 2009(6). Ren, Ruoen, Jie Li, Haitao Zheng, and Manying Bai, International Comparison of Economic Scales Between China and Japan, The Journal of World Economy, 2006(8). Ren, Ruoen, Haitao Zheng, and Manying Bai, International Comparison of the Economy Sizes of China and the USA, China Economic Quarterly, 2008(1). Wang, Lei, and Jing Zhou, Estimation of the Relative Price Level of China’s Provincial-Level Regions—Based on Generalized Space CPD Model, Statistics & Information Forum, 2012(8). Yu, Fangdong, A Study on the Methods, Results and Problems of the World Bank’s Estimation of China’s Purchasing Power Parity, Management World, 2008(5). Yu, Fangdong, The Methodological Improvement of 2011 Round of International Comparison Program (ICP), Statistical Research, 2011(1). Yuan, Wei, Dong Qiu, Ruoen Ren, Shantong Li, and Xinhua He. The Reviews of Some Experts on 2005 ICP Results by World Bank, Statistical Research, 2008(6).

4  VIEWING AND APPLYING THE PPP CORRECTLY 

95

English References Asea, Patrick K., and Enrique G. Mendoza, The Balassa-Samuelson Model: A General-Equilibrium Appraisal, Review of International Economics,1994(2). Balassa, B., The Purchasing-Power Doctrine: A Reappraisal, Journal of Political Economy, 1964(72). Cassel, G., Money and Foreign Exchange After 1914, The Macmillan Company, New York, 1922. Deaton, Angus, and Olivier Dupriez, Spatial Price Differences Within Large Countries, Princeton University and World Bank Working Paper, July 2011. Dornbusch, Rudiger, Expectations and Exchange Rate Dynamics, Journal of Political Economy, 1976(6). Froot, Kenneth A., and Kenneth Rogoff, The EMS, the EMU, and the Transition to a Common Currency, NBER Working Papers 3684, National Bureau of Economic Research. Hsieh, David, The Determination of the Real Exchange Rate: The Productivity Approach. Journal of International Economics, 1982(12). Isard, P., Exchange Rate Economics, Cambridge University Press, 1995. Lee, F., and J. Tang, Productivity Levels and International Competitiveness Between Canadian and U.S. Industries. American Economic Review, 2000(2). Obstfeld, Maurice, and Kenneth Rogoff, Global Imbalances and the Financial Crisis: Products of Common Causes, Working Paper, 2009. http://elsa. berkeley.edu/~obstfeld/santabarbara.pdf. Officer, Lawrence H., The Purchasing-Power-Parity Theory of Exchange Rates: A Review Article, International Monetary Fund Staff Papers, March 1976(23). Paige, D., and G. Bombach, A Comparison of National Output and Productivity of the United Kingdom and United States, OEEC, Paris, 1959. Pilat, D., The Economics of Rapid Growth: The Experience of Japan and Korea, Edward Elgar, Aldershot, 1994. Rogoff, Kenneth, The Purchasing Power Parity Puzzle, Journal of Economic Literature, 1996(34). World Bank, Measuring the Real Size of the World Economy: The Framework, Methodology, and Results of the International Comparison Program, World Bank, 2013.

CHAPTER 5

Understanding PPP Through Examples

5.1  Simple Examples of GDP Calculation Professor Perkins explains the calculation principle of PPP by presenting a simple example in his book Economics of Development.1 In this example, Professor Perkins assumes two economies (the United States and India), one commodity (steel products), and one element (employees in retail business). The steel products are tradable but not the employees. By producing one million tons of steel and charging $200 per ton, the United States has a total output of $200 million. In contrast, by producing 0.08 million tons of steel and charging 6000 INR per ton, India has a total output of 480 million INR. By assuming the absence of any transaction cost or trade barrier, the exchange rate is represented by the relative price of steel produced by these countries. Given that one ton of steel in the United States and India is $200 and 6000 INR respectively, the purchasing power of $1 is equivalent to that of 30 INR, thus representing the exchange rate. The quality of steel produced by these countries must be assumed as the same, expressed as “homogeneity” in statistical term. The following is another example. By providing 20,000 laborers in the service industry and charging $5000 per laborer, the United States has a total output of $100 million. Conversely, by providing 40,000 laborers and charging 30,000 INR per laborer, India has a total output of 1.2 billion INR. Based on the exchange rate (1:30), the salary of a 1 Gilles

Perkins, et al., Economics of Development, New York: W. W. Norton, 1996.

© The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_5

97

98  X. SONG Table 5.1  Case of purchasing power parity USA

Steel

India

Quantity (Q)

Price (P)

Production value (V)

Quantity (Q)

Q10

P10

V10

Q11

80,000 tons

6000 rupee/ ton

Q20

P20

V20

Q21

1 million tons

Service labor GDP

20,000

200 USD/ ton

5000 USD/ person 300 million USD

200 million USD 100 million USD

40,000

Price (P)

Production value (V)

P11

V11

P21

V21

30,000 rupee/person 1.68 billion INR

480 million INR 1.2 billion INR

Data Source Gilles Perkins, et al., Economics of Development, New York: W. W. Norton, 1996, p. 39

laborer in India in the local currency is converted to $1000, which is only 1/5 of the salary of laborers in the United States. As labor is untradeable, the relative price of labor between the two countries (1:60) does not equal to the exchange rate (see Table 5.1). Superscripts 0 and 1 refer to the United States and India, respectively, and subscripts 1 and 2 refer to steel and labor, respectively. To facilitate the comparison among many countries, countries with superscripts of 0 are chosen as the benchmark. If several economies participate in the comparison, their superscripts are set to j = 1, 2, · · · , k. The subscript is set to commodities or services that are included in the comparison. If n commodities or services are included in the comparison, they are given subscripts of i = 1, 2, · · · , n. Following the exchange rate method, we obtain the following:

GDP0 = V10 + V20 = P10 Q10 + P20 Q20 = (100 × 200) + (2 × 5000) = 30000(USD) GDP1 = V11 + V21 = P11 Q11 + P21 Q21 = (8 × 6000) + (4 × 30000) = 168000(INR) In the local currency, the GDP of the United States is $300 million and that of India is 1680 million INR. Based on the exchange rate (1:30), the GDP of India is converted to $56 million, and the GDP of the United States is 5.36 times larger than that of India.

5  UNDERSTANDING PPP THROUGH EXAMPLES 

99

Following the same exchange rate, the unit price of steel from India is converted from 6000 INR per ton to $200 per ton, resulting in a total output of $16 million. As steel is tradable, such a conversion is reasonable. By contrast, labor in India cost 30,000 INR per unit, which is converted to $1000 according to the exchange rate. Assuming that each unit of laborer in India has the same quality as that in the United States, both countries charge $5000 per laborer. In this way, a laborer in India is considered tradable. However, in reality, Indian labor is cheap and cannot be exported. Dividing the total value of Indian labor by the exchange rate underestimates the GDP of India. According to the PPP method, homogeneous commodities should have the same price. Both the commodities and services of the two countries participating in the horizontal comparison should have the same price. To facilitate the comparison among many countries, US prices are generally adopted as the benchmark, and the quantities of the two products or components from India are multiplied by US prices. In this way, the products of two countries are calculated at the same price, thus eliminating errors from the price difference.

GDP1 = P10 Q11 + P20 Q21 = (8 × 200) + (4 × 5000) = 21600(USD) Following the above formula, the GDP of India is $216 million, and the GDP of the United States is only 1.39 times that of India. As the same product or element must have the same price in the horizontal comparison, the PPP method is more reasonable than the exchange rate method. For India, the error between the results that are calculated by the PPP and the exchange rate method is 21,600/5600 = 3.86. Therefore, the GDP of India is underestimated 3.86 times by the exchange rate method.

5.2  Calculation Rules of the PPP Method The PPP method multiplies the relative prices by the domestic values of various commodities to obtain the GDP. In the above case, the GDP of India is computed by multiplying the price vector of the United States (P10, P20) by the output vector of India (V11, V21).

GDP1 = P10 Q11 + P20 Q21 = (P10 /P11 )P11 Q11 + (P20 /P21 )P21 Q21 = (P10 /P11 )V11 + (P20 /P21 )V21

100  X. SONG

where V11 and V21 are the outputs of the steel and service industries that are calculated in Indian rupee, respectively. The data are collected from national statistical agencies to ensure reliability. (P10/P11) and (P20/P21) are the price ratios of steel to the service industry using the United States as the reference country. In vector form, the GDP is calculated as follows:   j GDPj = Pi0 /Pi V j , j = 1, 2, · · · , k; i = 1, 2, · · · n where the subscript i refers to the commodities or services included in the comparison, the superscript 0 refers to the reference country (the United States), and the superscript j refers to the countries or regions participating in the horizontal comparison. A total of k countries or regions participate in the comparison, and each country or region has n commodities or services for comparison. Given that the US dollar is the international reserve currency, people usually take the United States as reference. However, the Asian Development Bank takes Hong Kong as the reference economy when comparing Asian economies. Mathematically, the above expression does not have any problem. If both economies only have one commodity and service as shown in the example and if no difference is observed between the commodities and services of these countries in terms of quality, the GDP (in US dollars) is computed by multiplying the relative price between these two countries by their domestic output. By calculating the GDP in this way, the economic sizes of different countries can be ranked.

5.3   Problems of the PPP Method In theory, the conversion described above is rigorous without any problems. However, the PPP method faces several challenges. First, each country utilizes a unique caliber for calculating the GDP. In the above formula, i represents the different commodities or services selected from the statistics. However, hundreds of thousands of commodities and services are produced and offered in the world, and there are more than 20 varieties for steel products alone. Therefore, taking all products into account when computing the GDP is impossible. Generally, only the bestselling commodities in the past years are considered in the GDP computation, and commodities without significant sales are ignored. The comparison of the purchasing power of

5  UNDERSTANDING PPP THROUGH EXAMPLES 

101

different currencies between countries should cover all trading objects in the market. However, the commodity intersections greatly vary between countries, that is, certain commodities are common in some countries yet rare in others. Only the commodities that can be found in both countries are comparable. Given that the economic structure of developed countries is different from that of developing countries, their consumption structures also differ. Therefore, the commodities that are included in the statistics basket are also different, and this difference leads to statistical bias. Second, the relative prices are unidentified. To calculate the PPP, the relative prices of various statistical commodities between two countries must be initially identified. If country A has a product with great output but country B does not have such a product, a relative price does not exist. To solve this contradiction, the International Comparison Program attempts to divide commodities and services into small groups. A representative commodity is selected in each group to calculate the price ratios of the common commodities. The results are then applied to the commodities that are rarely found in some countries. The ICP classification procedure of the UN is divided into the following stages: 1. All expenditures are grouped into basic heading, class, group, and category. 2. Some representatives from the most basic headings are selected, and the relative prices (price ratio) of the representatives between countries in the most basic headings are calculated. 3. The price ratio of the category with the price ratios of the representatives is represented. 4. The weighted average of the price ratios of the representatives is computed according to the shares of these products in the GDP to determine the PPP of a higher-level class. 5. Calculate the weighted average level by level until the total price ratio of all expenditures is obtained. However, this calculation method has several errors. If the representatives of two countries can cover the majority of the products of a category, the emerging statistical errors will be within the scope of permit. In fact, developed countries face a similar situation when selecting the category, and this procedure may only generate very small errors. However, some categories may be absent in developing countries. For example,

102  X. SONG

developing countries may have no high-tech products, and some primitive labor-intensive products in developing countries no longer exist in developed countries. The IT industry is very developed in the United States with high output. Conversely, many Asian and African countries have no computer software industry. In this case, how can we determine the price of the IT industry in developing countries and how can we calculate the GDP by using the PPP method without the relative price? Whereas Panama has a rich supply of bananas, Canada has a limited supply of this commodity because of the cold weather. Therefore, banana has no price ratio between these countries. Is it feasible to divide agricultural products into several groups, including a fruit group, and then select apple and its relative price as the representative of the whole group? The situation in Panama is not suitable for growing apples. Even if few farmers in Panama have grown apples, these fruits are mostly of poor quality and have a higher price relative to the other products in the country. Therefore, calculating the relative price of all fruits with the domestic prices of apples in two countries is a ridiculous attempt, let alone calculating the value of bananas with the relative price. In addition, Canada produces excellent potash ore, but Panama hardly produces potash fertilizer. In this case, errors will occur if the relative price of nitrogen is used to deal with the potash outputs of these two countries. However, some commodities, even those in small categories, have no intersection between developed and developing countries. As the connotations of many categories are different between poor and rich countries, using the relative prices of some commodities with intersections to represent the whole intersection will result in errors. A greater difference between two economies in terms of their economic development will produce a greater amount of errors in the PPP comparison. The ICP assumes that the price ratio of the specifications can represent that of the whole category, but the feasibility of representing the price ratio of non-specifications with that of the specifications remains a significant problem. If many commodities have no intersection between two countries and account for a large proportion of the products, representing the price ratio of the whole category with that of the specifications may result in huge deviations. Reflecting the purchasing power of a country with a biased sample can also lead to large deviations. Given the inherent transitivity of ICP in the calculation process, the biased sample also affects the purchasing power of other countries and reduces the accuracy of the whole system.

5  UNDERSTANDING PPP THROUGH EXAMPLES 

103

Third, the homogeneity condition cannot be satisfied. Only homogeneous commodities produced by two countries can be used for horizontal comparison. The PPP calculation formula tries to eliminate the between-country deviations that result from price difference by multiplying the commodity quantity by a unified price vector to represent the GDP. However, the commodity quantity does not contain the quality factor. For example, with statistics on annual car production of a country, the average price of cars is determined by dividing the total sales by the quantity of cars. However, the money spent on buying a Benz or a BMW can buy dozens of Geely and Chery cars. Moreover, the money spent on buying a large-screen HD TV can buy a dozen ordinary color TVs. By classifying these commodities into small groups, the deviations can be eliminated. However, many other problems, such as loss of statistical data and high cost, may emerge during the fine classification of commodities according to quality. Even commodities of the same quality may be sold at different prices because of their varying after-sales services. For example, TVs with different warranty periods are sold at various prices. Commodities that can be returned unconditionally are expensive. Therefore, including commodities with significantly different quality in the comparison can lead to illogical conclusions. Prof. Qiu Dong illustrates this issue by presenting a case in which several blind people feel an elephant. Some say that the elephant is a wall, and others say that the elephant is a rope. If the rope is taken as the specification, the deduced conclusion deviates far from reality because different parts of the elephant have poor homogeneity. If these blind people feel a haystack, several straws may be selected as the specification to explain the problem because of the high homogeneity.

5.4   Do Not Take the Experience of One Point and Spread It to the Entire Area or Take a Part for the Whole The price reflects the scarcity of resources in a market economy. A rarer product has a higher price. For example, bananas are sold cheaply in Panama, labor in Bangladesh is cheap, but the prices for both banana and labor in Sweden are expensive. In general, the resources endowment of a country is fixed. If such an endowment cannot be changed, the relative prices between countries tend to be highly inconsistent.

104  X. SONG

Some people say that the purchasing power of RMB is overestimated by the exchange rate between RMB and the US dollar, whereas others consider that such purchasing power is underestimated. These arguments contradict each other. For example, as the McDonald’s hamburgers (Big Mac) sold worldwide are only slightly different from one another in terms of quality, the ICP often quotes a Big Mac index. A hamburger is sold at 25 CNY in China and $4 in the United States. Therefore, the relative price (PPP) is 6.25 (25/4), that is, the purchasing power of $1 is equivalent to that of 6.25 CNY. The exchange rate of the US dollar against the RMB is equivalent to 1:6.20 in 2014, which indicates that the PPP is equal to the exchange rate. By dividing the relative price by the exchange rate, the PLI is 1.01 (6.25/6.20). Given the high homogeneity of McDonald’s products, it makes sense to discuss purchasing power using the price of Big Mac. All commodities with high liquidity and fierce competition in the international market can form a unified market relatively easily. Therefore, the PPP of different currencies on these commodities can be easily computed. Each country has different resource endowments that are difficult to change in the short term. Scarce elements and commodities are normally sold at higher prices. Developing countries have a relatively low capital and a surplus labor force. As the labor force cannot flow across countries, a unified labor force market cannot be established worldwide. Therefore, the wage levels across countries also vary. A large amount of surplus labor still remains in the rural areas of China. In 2013, it was estimated that approximately 88 million rural surplus laborers exist in China. The average prices of many services in China are lower than those of services in the United States. In terms of in-person service, the purchasing power of 6.20 CNY is stronger than that of $1. Let us take hairdressing as an example. A simple haircut costs $30 in the United States and 20 CNY in China, thus amounting to a PPP of 0.67 (20/30). That is, the purchasing power of $1 is equivalent to that of 0.67 CNY. The PLI is 0.11 (0.67/6.20). Many young people in China prefer famous fashion brands, and thus the prices of foreign luxury products become expensive. Surprisingly, the purchasing power of the US dollar is very high in some luxury product markets. For example, a Louis Vuitton bag sold at $200 in the United States may be sold at 5000 CNY in China, with a PPP of 25 (5000//200). In other words, the purchasing power of $1 is almost 25

5  UNDERSTANDING PPP THROUGH EXAMPLES 

105

times greater than that of 1 CNY. Therefore, the PLI is 4.03 (25/6.20). The high prices of luxury products in China are also associated with “grey” income. As many imported luxury products are not considered in the GDP calculation, the low weight and the pushing effects of the prices of luxury goods on the overall price of China are not reflected. The local reliability of the PPP method can be demonstrated through several more examples. According to the National Bureau of Statistics of China, the per capita disposable income of urban households in 2011 amounted to 21,809.8 CNY, in which the per capita food consumption expenditure accounted for 5506.3 CNY (25.25%). In the same year, the per capita net income of rural households amounted to 6977.3 CNY, in which the per capita food consumption expenditure accounted for 2107.3 CNY (30.20%). However, the results of ICP 2011 showed that the actual expenditure on food and non-alcoholic beverages in China was $5.8115 trillion and the GDP of China was $13.4959 trillion, with the former accounting for 43.06% of the latter. Clearly, the proportion of food consumption was overestimated. On the other hand, many catering services are included in food consumption. As food prices are lower than the prices of catering services, the overestimated weight magnifies the price deviation. This mismatch eventually decreases the overall price. Let us take public consumption as another example. Price is computed as cost times the markup percentage. The price ratio can be decomposed into cost ratio and ratio of markup percentages. Developed countries have higher cost, whereas China has a higher markup percentage because of corruption and other efficiency problems. At the same time, the high centralization of China results in a high decision-making efficiency. Therefore, the final markup percentage of China is not a lot higher than that of developed countries. In other words, the price of public consumption in developed countries is higher than that in China. However, such differences in public consumption are mainly reflected by the high cost in these countries. The actual expenditure cannot be computed by using cost as the conversion factor because a high cost may also result in borrowing. As a high public consumption cost effectively shows the strength of a country (in contrast to a high food production cost), breaking down the expenditure is necessary when investigating international consumption expenditure but is unnecessary when investigating the domestic consumption expenditure of a country.

106  X. SONG

Many more examples can demonstrate the contradictions discussed above. All these examples prove that it is extremely difficult to maintain the homogeneity of samples when the PPP is used for calculation. In the absence of homogeneity, a simple aggregation of data for horizontal comparison will inevitably produce serious deviations.

CHAPTER 6

Calculating China’s GDP

6.1  Two Types of Accounting Systems 6.1.1   Four Accounting Systems The UN Statistical Commission cooperates with the World Bank, the OECD, the EC, and the IMF in detailing the following four accounting systems: the System of National Accounts (SNA), the System of Material Product Balances (MPS), the System of Social Demographic Statistics (SSDS), and the System of Integrated Environmental and Economic Accounting (SEEA). These systems are used in guiding economic accounting, material products accounting, social population accounting, and environmental accounting of different countries worldwide. Among the four accounting systems, two norms are related to national accounting, namely, MPS and SNS. 6.1.2   System of Material Product Balance (MPS) MPS originated from the Statute of State Statistics of the Soviet Union signed by Lenin in 1918 that compiled the balance sheets of grain and feed and proposed to compile the balance sheet of national economy. Later, the Soviet Union’s Central Bureau of Statistics issued a series of national economy balances and compiled the input–output sheet based on the experience of SNA 1953. The balance sheets were recognized by international organizations. The UN Statistical Commission released © The Author(s) 2019 X. Song, Understanding Chinese GDP, https://doi.org/10.1007/978-981-32-9733-3_6

107

108  X. SONG

the Basic Principles of the System of Balances of the National Economy as an official document in 1971, which included four main balance sheets and 13 supplementary sheets. In 1977, it published the Comparison of the System of National Account and the System of Balances of the National Economy. In 1984, the standing statistical committee of the Council for Mutual Economic Assistance, with the Soviet Union and the Eastern European socialist countries as the main members, made major revisions and supplements of MPS to form the so-called new MPS. The council published the Basic Methods and Principles of Compiling the Balance of National Economy Statistics, which added the balances of non-material services and interdepartmental balances, household income, and consumption index. MPS divides all the departments of the entire society’s economic activities into two fields: material production and non-material production. Based on this classification, MPS forms two core indices throughout the entire accounting system, namely, total social output value and national income (also known as the net social output value). Total social output value is defined as the sum of the material products’ value produced by all material production departments of the entire society within a certain period of time. National income is defined as the sum of new values created by laborers in the material production field of a country (region) within a certain period of time. The relationship between them is expressed as follows: national income  =  total social output value − (intermediate consumption value of the material production departments + depreciation of fixed assets). MPS mainly reflects physical movements in the production, exchange, and use of material products but ignores capital flow accounting, national wealth accounting, and service production of non-material production departments. At the same time, MPS adopts the single-entry bookkeeping method to unilaterally enter and calculate the national economic activities. One of its important accounting tools, the balance sheet, focuses only on internal balance. Despite the connection among various balance sheets of the national economy in terms of quantity, they lack a strict calculation relationship among one another. MPS is mainly used in the national accounting of planned economy countries. With the development of the global market economy since the 1990s, planned economy countries have shifted to the market economy. Consequently, MPS has pulled out of the national accounting practice and has become a thing of the past.

6  CALCULATING CHINA’S GDP 

109

6.1.3   System of National Accounts (SNA) After the Second World War, SNA gradually became the most widelyused accounting system. In 1950, the UN Statistical Office extracted the national income statistics in 1938 and 1948 from the original data of 41 countries, including 13 sets of accounts or accounting tables compiled by the countries using the social matrix method of accounting. That same year, the Organization for European Economic Cooperation (OEEC) released an account system that formed the basis of the SNA’s standardized version published in 1952. By 1953, an expert team from the national economic accounting research approved by the SecretaryGeneral of the UN held a meeting in New York to launch A System of National Accounts and Supporting Tables (1953 report). This system defined the classification of three basic institutional sectors (enterprises, residents, and private non-profit organizations), including six standard accounts and 12 standard tables. This version became the first version of the Systems of National Accounts and marked the birth of a unified SNA. SNA 1953 aimed to provide an accounting standard for measuring the national income for all countries worldwide. This standard was generally applicable. Its important features were as follows: considering the needs of developing countries for national accounting, designing the agricultural sector balance sheet, providing independent accounts for the non-monetary transactions of developing countries, and emphasizing the importance of maintaining consistency in international statistical standards. Since then, as a general norm of the national accounting and international statistical data comparison, SNA 1953 had been gradually adopted by many countries worldwide. Objectively, no norm and standard could be widely applicable to all countries. SNA 1953 faced various challenges at the beginning of its application. In practice, the UN experts continued to revise and improve upon it. After 15 years, the UN Statistical Commission approved the revised SNA based on a paper by Rechard Stone and the literature selected by the OECD, which is the second version of SNA called SNA 1968. SNA was revised twice in 1960 and in 1964 between SNA 1953 and SNA 1968. The first revision reconciled the deficiencies and defects of SNA after its launch in practice and expressed the willingness to establish the international standards and expand the scope of accounting. The second revision adjusted some accounting contents and corrected some errors based on the IMF’s Balance of Payments Manual.

110  X. SONG

In the 25 years between 1968 and 1993, the UN was committed to the SNA’s revision and improvement, constantly requiring countries to provide feedback to the UN on the progress and problems encountered in a timely manner. In 1975, the UN began summarizing the different countries’ experience in SNA practice again. In 1980, the UN expert group discussed the application situation and future development direction of SNA according to the experience of different countries and submitted a report to the statistical commission. The report emphasized the status and role of SNA as a standard system. Between 1982 and 1985, the expert group made minor revisions to the SNA. In 1986, the expert group held a meeting to discuss SNA’s structure, price and quantity comparison, foreign department, household sector, public sector, production accounts, investment and output table, financial flow, and integration of MPS and SNA. In 1989, the UN established an expert group to review the drafts of all revised chapters of SNA and to solve prominent problems. After six meetings, the expert group submitted a tentative revised SNA draft to the statistical commission in 1991. In 1992, the expert group again improved and reorganized the contents of all chapters in the draft and its appendix and submitted the discussion results to the statistical commission along with the SNA revision in 1993. The experts had claimed that this version of SNA had made great improvements to SNA 1968. The UN Statistical Commission passed a resolution to accept this version, that is, SNA 1993. Simultaneously, the UN Economic and Social Council widely recommended SNA 1993 to its member countries and relevant international organizations. SNA 1993 served as an instrumental document of national accounting theory because of its rich content, large system, rigorous structure, and scientific method. It represented a new stage of SNA’s development and was regarded as the standard by scholars and practitioners in the field of national accounting in all countries worldwide. Many countries developed SNAs that suit their national conditions using SNA 1993. However, as time progressed, SNA 1993 failed to keep up with the times. In 2003, the UN Statistical Commission decided to update SNA 1993 to better meet the demands of national economic frameworks and data users. This updating work lasted until 2008. With the support of the national accounts secretariat working group, the updated SNA was reported to the UN Statistical Commission in two volumes and was approved at the 39th and 40th sessions of the UN Statistical Commission in 2008 and 2009, respectively. Finally, the two volumes

6  CALCULATING CHINA’S GDP 

111

were combined as a unified document called SNA 2008 to be officially released in 2009. SNA 2008 contains 29 chapters, over 730 pages, and over 140 accounting sheets. The contents can be roughly divided into three parts: (1) concepts, principles, and rules of national accounting, (2) account system of national accounting, and (3) accounting of relevant economic activities. This version of SNA creates a step forward in accounting content, scope, and methods compared with previous versions. Particularly, it notably focuses more on the connection between environmental resources and economic activities, introduces the satellite account and Social Accounting Matrix (SAM) for environmental-economic accounting, and provides a more detailed description of the compilation of environmental and resource accounts. After the official release of SNA 2008, developed countries with advanced statistical capacity and more comprehensive databases have begun to implement or plan to implement the new international standard. Some countries estimated its influence on the statistical results of their own Gross Domestic Products (GDPs) according to the changes in the accounting rules. For example, the United States estimated that its GDP in 2012 would increase by 3.6% according to the statistical rules of the latest NIPAs. Canada released its revised GDP data according to the latest accounting system in 2012, with its annual GDP increasing by about 2.4% on average between 2007 and 2011. Australia began to consider R&D expenditure as a fixed asset in 2009. Consequently, its GDP in 2008 increased by about 1.45%. Although the EU countries did not officially implement SNA 2008 until September 2014, the Netherlands had begun the GDP revision work before that. According to the released results, its GDP in 2010 increased by 3.0% with the use of the latest rules. Japan planned to implement SNA 2008 in 2016.

6.2  Evolution of China’s National Accounting 6.2.1   Recovery Period of the National Accounting (1978–1984) The development of China’s SNA can be divided into three stages. During the recovery period of China’s national economy in 1951, to meet the needs of the country for the balance between materials and supplies and between revenue and expenditures, relevant departments made

112  X. SONG

exploratory calculations of the financial and material resource balances, established the balance of agricultural products, industrial production goods and consumer goods, and then expanded the product varieties of those balances. In 1952, the statistical agency of the new Chinese government conducted an investigation of gross industrial and agricultural output value and labor employment since the early days of the PRC’s founding. It slowly expanded the accounting of gross industrial and agricultural output value into the accounting of gross output value of five material production departments (industry, agriculture, construction, transportation, and commerce) to reflect the scale, structure, and speed of the national economic development. Since 1954, the National Bureau of Statistics has begun performing the production, distribution, consumption, and accumulation accounting of the national income based on study of the Soviet Union’s statistical theory and method of national income. It provided a series of national economic aggregates and national income accumulation rates that served as an important basis for understanding China’s national economic development and strengthening the planning and management of the national economy. In 1956, the National Bureau of Statistics sent a delegation to conduct an overall investigation on the Soviet Union’s national accounting work and then implemented MPS in China. It compiled the balance of production, accumulation, and consumption of social products; the balance of production, distribution, and redistribution of social products and national income; the national economic sector interdepartmental balance; labor resources; and the distribution balance. MPS was appropriate for the highly concentrated planning management system in the early days of the new China, and it made important contributions to the development of the national economic plan and macroeconomic management. China’s statistical work mainly followed MPS until 1984, in the early stage of the reform and opening up. Despite China’s adoption of MPS, because of the restriction on various conditions, China did not completely implement MPS and only applied some parts of it in accordance with the actual demands. Therefore, China’s national accounting system was neither systematic nor complete at this stage. Particularly after the reform and opening up begun in the late 1970s, the accounting data could neither completely reflect the landscape of China’s national economy and the composition of industrial departments nor meet the demands of international comparison due to the limitations of the MPS and to the progress of

6  CALCULATING CHINA’S GDP 

113

China’s economic system reform. With the demand for macro-management and the needs of diplomacy, establishing a new national accounting system that had international comparability and that could completely reflect the national economy’s landscape was necessary. 6.2.2   Reform Period of National Accounting (1985–1992) Adapting to the new situation after the reform and opening up compelled the National Bureau of Statistics to start establishing a new system of national accounting in 1984 while continuing the implementation of the MPS. In January 1984, the State Council proposed the establishment of a unified and scientific national accounting system in the Decision on Strengthening the Statistical Work. In April 1985, the State Council approved the establishment of the GNP and the third industry statistics. Furthermore, the National Bureau of Statistics began ­making GDP calculations based on the national income. Toward the end of 1984, the State Council set up a steering committee that would lead the work on establishing accounting standards for the national economy. Under the leadership of this group, the National Bureau of Statistics and the relevant departments conducted a series of tasks on the construction of a new national accounting system based on previous practice and theoretical research. It developed China’s System of National Accounting (Trial) in August 1992. The trial method still held MPS in a relatively important position and displayed a tendency toward integration. It considered maintaining the historical variability and macroeconomic analysis of China’s national accounting information, the habits of managers, and the international comparability of the national accounting information of countries that were still using MPS at that time. 6.2.3   Development Period of National Accounting (Since 1993) The following three reasons contributed to China’s move away from MPS. First, the socialist market economy theory challenged the theoretical basis for the MPS. The 14th National Congress held in October 1992 established the reform goal to construct the socialist market economy system and realized a breakthrough of socialist economic theory, which cleared the theoretical obstacles for the comprehensive reform and development of China’s national accounting.

114  X. SONG

Second, meeting the needs of China’s macroeconomic management was difficult for MPS. The practice of China’s national accounting work showed that the deficiencies and defects of the MPS had become apparent in reflecting the changes in the national economic development. The definition of production scope in MPS was too narrow to fully reflect the economic status of the different types of market entities, including their mutual relations and interaction. The accounting principles of the MPS failed to fully reflect the characteristics of the market system and open economy. The accounting method was too singular that it could not meet the needs of the socialist market economic system. Third, the international comparability and generalizability of the MPS gradually weakened. Due to the reform in politics, society, and system, the Soviet Union--where MPS was originally conceived--and Eastern European countries, abolished the MPS and changed to SNA in the early 1990s. In 1993, the 27th session of the UN Statistical Commission passed a resolution to cancel the MPS and globally use the SNA instead. In 1993, adapting to the changing situation compelled China to cancel its national accounting under the MPS. In 1995, China began preparing the asset debt balance sheets and national economy account using SNA. In 1993, the statistical department of the Chinese government made significant revisions to the China’s System of National Accounting (Trial) of 1992 in accordance with the standards of SNA 1993 while developing China’s System of National Accounting (2002). As much as it could, it considered both the integration of the international standards and China’s actual situation to make the structure more rigorous and to enrich the content. China’s SNS covered the main links and aspects of the national economic operation under market economic conditions and fully reflected the internal connection of national economic activities. This system thus was more feasible to implement. Furthermore, it could better adapt to the needs of macroeconomic management and foreign exchange under the socialist market economy. Presently, an increasing number of countries worldwide are implementing SNA (2008). To synchronize with international norms, China would implement a new SNA toward the end of 2014 or in early 2015. Due to its weak basis of national accounting (compared with the latest international standard and that of developed countries), China’s SNA still has gaps and requires continuous development and improvement in practice.

6  CALCULATING CHINA’S GDP 

115

6.3  Major Adjustments of the Official Statistical Data Throughout history, China’s GDP accounting has had two major supplements and three major adjustments. 6.3.1   Two Major Supplements China’s GDP accounting began in 1985. To satisfy the requirements of the macroeconomic analysis and management of data continuity and comparability, China added two major supplements on the historical data of its GDP. The first supplement done between 1986 and1988 amended GDP calculations for 1978–1984, a period after initiation of the reform and opening up. The second supplement done between 1988 and 1997 amended GDP calculations for 1952–1977, a period before the reform and opening up. The contents of the two major supplements were basically the same, including both the GDP production accounting and the GDP application accounting. Their methods were also the same. In terms of production accounting, the supplement first adjusted the net output values of the five material production departments, namely, agriculture, industry, construction, transport and telecommunications, as well as wholesale and retail trade and catering. It deducted the payments for non-material services (e.g., financial insurance service fees, advertising and information, consulting fees, etc.), added the depreciation of the fixed assets to obtain the added values of these departments, calculated the added values of non-material production departments, and finally summed up the added values of the material production departments and non-material production departments to obtain the GDP. In terms of application accounting, the final consumption and capital formation (accumulation), including goods and service import and export, were supplemented and adjusted in the national income. The supplement and adjustment of the final consumption is also the process of deducting the residents’ consumption and the material product values consumed by the non-material production departments in government consumption from the national income. Furthermore, it added all the expenditures of the residents and the government for the services provided by the non-material production departments to form residents’ consumption and government consumption of the GDP. The supplement and adjustment of capital formation was intended to add

116  X. SONG

the depreciation of fixed assets to the net fixed capital formation (fixed assets accumulation) of the national income to form the fixed gross capital formation in the GDP. The supplement and adjustment of goods and service import and export was intended to supplement the non-material service import and export into the goods and service import and export of the national income, forming the goods and service import and export of the GDP. The first batch of supplementary data was briefly published in the China Statistical Yearbook in 1988 and the second batch was published in the book History of China’s GDP Accounting (1952–1995). The book published the detailed supplementary data of the first batch. 6.3.2   Three Major Adjustments The first major adjustment of GDP data was made in 1994 and 1995 after the first census of tertiary industries in China. The second and third adjustments were conducted after the first and second economic censuses in China in 2004 and 2008, respectively. The long-term emphasis on the production of material products and the chronic use of the MPS resulted in the failure of China to attach due importance to the statistics of production activities in non-material service industries. After the implementation of GDP accounting in 1985, the data source of the non-material service production activities had always been a weak link. Non-public wholesale and retail trade, catering, and transport industries had achieved rapid development during the reform and opening up era, but conventional statistics failed to cover all these economic activities. To solve these contradictions, China implemented the first census of tertiary industries in 1993 and 1995 (census years: 1991 and 1992). The first major adjustment on the historical data of the GDP was made according to the information obtained from the census. The time scope of the adjustment was 1978–1993, a total of 16 years. The adjustment contents included the production accounting and the application accounting of the GDP. The adjustment of production accounting included the adjustment of the added values of all industrial departments in the tertiary industries and the GDP. The adjustment of application accounting focused on the adjustment of the final consumption and the GDP calculated by the expenditure method. The GDP and its structural data in production and application after the first major adjustment were first briefly published in the China

6  CALCULATING CHINA’S GDP 

117

Statistical Yearbook in 1995, and the detailed data were published in the History of China’s GDP Accounting (1952 to 1995). The two major supplements of the historical data of the GDP promptly satisfied the needs of macroeconomic analysis and management for the corresponding data. The first major adjustment enabled the GDP’s historical data to reflect more accurately the development of tertiary industries and formed a more solid basis for the country’s establishment of reasonable industrial policies. In 2003, the National Bureau of Statistics adopted two major reform measures on China’s statistical system. First, it combined the original industrial census, the census of the tertiary industry, and the national basic unit census. At the same time, the construction industry was integrated into the national economic census and would be done once every five years. China’s periodic national census system previously included a population census, agricultural census, industrial census, census of the tertiary industry, and basic unit census. After adjustment, the national censuses were reduced to a population census, agricultural census, and economic census. The first national economic census began in 2004, and its targets were legal entities, industry units, and individual industrial and commercial units engaged in second and tertiary industries in China. Since then, the national economic census has been conducted once every five years. Second, the GDP accounting system and the data release procedure were improved while the process for regular revision and adjustment matured. The GDP accounting and data release procedures after the reform were: The forecasts of the year were no longer released at the end of the very year and preliminary accounting data would be released on January 20 of the following year instead of February and May of the following year. The accounting and release procedures of the preliminary and final validation data remained unchanged. In the future, a revision and a release of the historical GDP data would be mandatory when new resources, methodology, and classification standards became available. Both the total GDP and the growth rate are to be revised accordingly. After the first economic census, revising the historical data in accordance with international practice became necessary to keep historical comparability of the GDP data. With the GDP accounting data of the economic census year 2004 as the basis, the National Bureau of Statistics adopted the trend-deviation method commonly used by the OECD to revise the historical macro-statistics data, including the GDP since 1993.

118  X. SONG

The revised data were first released in the China Statistical Yearbook published at the end of 2006. The change in the accounting method enabled the revision of the historical data to be traced back to 1953. The second economic census year was 2008. After the census, the National Bureau of Statistics again adjusted the previous annual data in accordance with the prevailing international practice. The time interval of this adjustment was between the first economic census year and the second economic census year, namely, between 2005 and 2008. The National Bureau of Statistics adjusted not only the GDP but also the added values of different industries, including the present price data and the constant price data. The standard time of the third economic census was December 31, 2012, and the census period was from January 1, 2013 to December 31, 2013. After the census, the National Bureau of Statistics made the corresponding adjustment to the data in 2009 and 2012 in accordance with the international practice. Considering that the economic census coincided with the release of China’s new national accounting system, the bureau further revised China’s historical data based on the census data, this time in accordance with the contents of the new accounting system.

6.4  The Sum of Regional GDPs Is Greater Than the National GDP 6.4.1   The Discrepancy Between the National GDP and the Sum of Regional GDPs Calculated by the Production Method The fact that the sum of regional GDPs is higher than the national GDP has plagued economists and officials in recent years. During the first half of 2014, the preliminary calculation amount of the national GDP reached 26.9044 trillion yuan, and the sum of GDPs of the 31 provinces, cities, and autonomous regions reached 30.283559 trillion yuan. The latter was 12.56% higher than the former. Why is the sum of regional GDPs greater than the national GDP? This question is worthy of further investigation. The GDP can be calculated using three methods, namely, the production method, the income method, and the expenditure method. In the non-census years, only the data calculated by the production method and the expenditure method are released. Thus, this chapter only analyzes the data calculated by the production method and the expenditure method.

6  CALCULATING CHINA’S GDP 

119

The production method aims to measure the new values created by the permanent units in a certain period of time from the perspective of production, namely, deducting the values of intermediate goods and services input from the production process to obtain the added values. The expenditure method aims to measure the final results of the production activities of all permanent units in a certain period of time from the perspective of final use. The final use includes the final consumption expenditure, the total capital formation, and the net export. In consideration of the availability of the data, the national GDPs and the sums of regional GDPs between 1990 and 2012 are selected for comparison. The specific changing trend is shown in Fig. 6.1. Figure 6.1 shows that before 2002, the national GDPs were more than the sum of regional GDPs, but the size of the gap was insignificant; after 2002, the sum of regional GDPs exceeded the national GDPs, and the gap grew wider. Figure 6.2 shows that the difference in 2007 was less than that of the previous year. After 2007, however, the difference demonstrated a sharp rising trend. The trend for the percentage change of the difference is complex. It peaked in 2003, and the percentages

Fig. 6.1  Trends of national GDP and the sum of regional GDPs between 1990 and 2012 (Source China Statistical Yearbook for respective years, National Bureau of Statistics of China)

+XQGUHGELOOLRQ&1