Galbraith China Hot Money

Is China Really Running a Trade Surplus? James K. Galbraith Professor Lyndon B. Johnson School of Public Affairs The University of Texas at Austin Austin, Texas Sara Hsu Adjunct Professor of Economics St. Edwards University Austin, Texas Li Jianjun Professor of Economics Central University of Finance and Economics Beijing, China Last updated: December 30, 2007 The University of Texas Inequality Project UTIP Working Paper 45

Abstract We examine China’s macroeconomic and trade accounts for simple, tell-tale signs that capital inflows are being disguised as export earnings. We find large reported increases in a calculated unit value of Chinese manufactured exports, which do not appear to correspond to increased unit prices in the accounts of countries importing from China. We therefore suspect that the legalization of dollar accounts by firms resident in China, as well as an increase in expectation of RMB appreciation which occurred in 2003, have led to large disguised capital inflows. The magnitudes could range up to $529 billion by 2006. If this is correct, then China is not running a $170 billion current account surplus as officially reported in 2006, but rather a much smaller surplus, or even a deficit, obscured and financed by illicit inflows.

1

Introduction One of the most basic principles of international macroeconomics is that the growth of imports depends on the domestic rate of growth, while that of exports depends on growth in external markets. For a country selling manufactured goods to the whole world, the relevant comparison is surely between the national growth rate and that of the world economy, or at least that of the OECD countries, which provide the lion’s share of the global consumer market. Thus when a developing country experiences a prolonged period of high growth of internal demand, it is normal for a trade deficit to emerge. This is especially likely if the country in question is an importer of food and fuel and commodity prices are rising. And if the country faces a finance constraint, the trade deficit will ultimately limit the growth surge. Innumerable cases can be cited, in Latin America, Africa, and even Europe. Exceptions, per contra, are rare, and in the modern record largely confined to countries that maintain rigorously undervalued exchange rates and repressed domestic consumption, while rapidly improving the composition and quality of their exports. Seen from this perspective, the recent record of the People’s Republic of China is simply astounding. China has been running reported internal real growth rates of eight percent or so for three decades, during which time OECD growth rates averaged less than half of that figure (WDI Online). And while during most of this period China reported small trade surpluses, in the most recent years China’s current account surplus has exploded. China’s reported exports nearly quadrupled from 2000 to 2006, from $249 to $969 billion, a rise of nearly three-quarters of a trillion dollars. Despite the fact that China’s imports rose substantially in both quantity and unit prices during this period, China reported a trade surplus of over $170 billion in 2006. But if this feat seems improbable, there is the possibility that it didn’t happen. In this paper, we show that a very large fraction of the gains in reported Chinese exports after 2002 are apparently due to rapidly rising unit values, following a long period in which unit export values did not rise at all. Only 40 percent, at most, of the export rise can be attributed firmly to rising export quantities. The question we pose is: did the per-unit values of Chinese manufactured exports really rise at rates exceeding 20 percent for three consecutive years? Looking at import prices for Chinese goods recorded in other countries and similar evidence, we find no trace of such a transformation in unit prices, and only modest shifts in the composition of exports that might be raising measured unit values. That being so, the possibility arises that Chinese exporters have been over-reporting export prices to the Chinese authorities, for the purpose of bringing foreign exchange into the country. The incentive to do so stems from two facts: the continued enforcement of strict controls over capital inflows per se in China, and the legalization, in late 2002, of unlimited foreign currency accounts held in China by Chinese firms. On the simple evidence of reported price increases, the disguised capital inflow could be very large: potentially enough to turn that reported surplus of $170 billion in 2006 into a substantial trade deficit, and enough to explain the very large increase in the share of fixed investment in Chinese GDP that occurred after 2002.

2

Capital Inflows into China Hot money has been flowing into and out of China since its emergence as a major economy. And the Chinese government, which has a long history of capital control, is concerned about the issue. Since capital controls remain in force in China while the current account was liberalized in 19961, it stands to reason that some efforts to evade capital control would flow through the current account. Officials at the People’s Bank of China (PBOC) have confirmed that significant hot money inflows have run through the current account2 through the over-invoicing of exports, the practice of overstating export values in order to bring foreign capital into the country. Hu Xiaolian, director of the State Administration of Foreign Exchange (SAFE), and vice governor of the PBOC, and Deng Xianhong, deputy head of SAFE, recently called for audits of short-term foreign exchange accounts to check these inflows (Xin 2007)3. An anonymous governor of PBOC, associated with the National Development and Reform Commission (NDRC), has estimated that "false exports during August-December 2006 resulted in an increase of US $17.5 billion in the favorable foreign trade balance, accounting for 17% of the total favorable balance” (Zhong 2007).4 It appears that the Chinese government has estimated the occurrence and extent of capital inflows in the trade account by watching short-term foreign exchange transactions, upon which the PBOC imposed further regulations in February of 20075, Scholars and bank economists have long watched Chinese hot money fluctuations, mainly via the catch-all errors and omissions category in the balance of payments (Prasad and Wei 2005). Green (2006) estimates that hot money inflows comprised US $67 billion in 2005, although this is a very rough estimate. Some have also suspected that hot money has also flowed into the current account, taking the form of payments for fictitious exports or over-invoicing of actual exports. This phenomenon appears not to have attracted extensive scrutiny so far.6 Yet, it could have large implications for understanding the true trade and financial position of China today In this paper, we investigate the extent to which capital inflows may have appeared deceptively in the trade account from 2003 to 2006. Not having access to the Chinese government’s financial data, we take a simple alternative approach: we estimate the inflows using the published balance of payments data, while checking our estimates against investment, foreign import prices, and financial activity. 1

See Li (2004). “Capital Account Liberalization in China,” The Chinese Economy, 37(1), pp. 85-116 for a timeline of current account liberalization. 2 Wei and Zhang (2007) note that a government official of an “anonymous” country admits that capital inflows have been introduced through the overbilling of exports. That the reference is to China is consistent with the context of this remark. 3 Nineteen domestic banks and ten international banks have been punished for facilitating short-term money inflows disguised as trade or investment (Anderlini, Financial Times, June 27, 2007). 4 On the other hand, Sun Mingchun, vice-president and Asia economist of Lehman Brothers Asia Ltd, stated that hot money inflows may be slowing down due to recently implemented checks on short-term capital inflows and stock market transactions (Zhang July 13, 2007). 5 See People’s Bank of China Adjusted Foreign Exchange Administration Policy towards Individuals 6 Gunter (2003) makes a case for the phenomenon of capital flight from China, from 1984 to 2001, with overinvoicing of imports.

3

China’s unlikely export values A simple first cut at the problem involves making a large, problematic assumption: that changes in the internal composition of Chinese exports within major product categories over a short three-year period can be safely ignored. (We will examine this assumption in detail later.) If we can treat the commodity composition within the major product categories as fixed, then dividing total export values by quantities will give us a “unit value” measure of Chinese exports.

The official data available for this purpose are very erratic, but it is easy to show that this is due mainly to the influence of two highly volatile export sectors, neither of which is quantitatively very important7. We therefore construct a streamlined representation of export volume by removing those sectors. For the remaining sectors, we find a steady increase in export volumes over time. Dividing dollar volume by these quantities gives unit values8. We calculate unit values using both OECD and CEIC data, applying the growth rate in calculated unit values after 2004 to quantities given in the OECD data9. Table One shows export unit values from 1996 to 2006 calculated after excluding the volatile sectors10. In manufacturing, these values are extremely stable through 2002, at around $0.39 USD per reported unit, and then they start increasing rapidly. In 2003, the manufacturing export unit value jumps to $0.49, and in 2004, to $0.59. In 2005, the export unit value is $0.72, peaking at $0.97 in 2006, and the overall export unit value follows the same trend, which is not surprising since manufactures dominate Chinese exports. Table One: Export Unit Value (In US Dollars) Export Unit Value (Our Calculations) Export Unit Value (Official Statistics) Manufacturing Export Unit Value (Our Calculations)

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

0.51

0.47

0.47

0.46

0.46

0.45

0.46

0.56

0.66

0.79

1.00

0.05

0.06

0.07

0.08

0.10

0.08

0.60

0.70

0.76

N/A

N/A

0.40

0.38

0.39

0.39

0.39

0.38

0.40

0.49

0.59

0.72

0.97

7

See Table Five below. The volatile sectors are Beverages and Tobacco (SITC 1) and the Special Commodities and Transactions (SITC 9) 8 This calculation is far from precise, because quantities are organized by type of unit, such as tons, thousands of units, and so forth, depending on sector, so there is no single consistent “unit of exports.” For this first cut, we are in effect assuming the existence of a constant “composite unit,” not strongly affected by changes in the composition of exports or by quality change. 9 OECD data is the only data set that contains total quantities for SITC categories and for all trade, but only goes up to 2004. CEIC makes some quantities available, but not all. Therefore, we needed to use both. 10 The World Bank’s WDI Online database includes information on China’s export value indices. These are also shown to increase dramatically from 2003 to 2006. We choose not to use this information because data from another source, UN Comtrade, do not show the same increase.

4

Source: OECD and CEIC Data and authors’ calculations11 Now, let us suppose that Chinese exports had continued at unchanged unit values for the entire period from 1997 through 2006. What would have been the growth of total exports on that assumption? Table Two gives actual export values, export values under the counterfactual of no unit value change, and the difference, which is attributable to changing unit values. It is clear that a major part of the reported increase in Chinese exports is not due to increasing raw volumes, but rather to some combination of reported price increases and product transformation, reported as rising unit value.

Table Two: Export Volumes and Total Exports Attributable to Volume Gain (in Billions of Current US Dollars)

2002 2003 2004 2005 2006

Exports of Goods and Services 365.4 485.0 655.8 836.9 981.0

Exports of Goods and Services at Unchanged Unit Values 365.4 398.4 457.1 490.2 452.4

Difference12 0 86.6 198.7 346.7 528.6

Source: World Bank WDI and authors’ calculations

As noted, part of the increase in unit values can be attributed to actual price increases in exports, and part to shifts in the composition of Chinese manufacturing output to higher technology goods. But how much? That is the question we next examine. Have Chinese manufactured exports increased in unit price, that is, in value per item holding composition and quality constant? If they had, we would expect that the unit import prices in manufactured goods reported by other countries, especially Europe, Japan and the United States, would show comparable increases. However, as Tables 16 to 18 in the appendix demonstrate, 11

Rather than using given total quantities, the total of the individual commodity categories (minus the extremely volatile beverages and tobacco category) was used, since the two were not equal. Using given total quantities would result in an even more dramatic increase in value per exported unit. The OECD does not yet have data for 2005-6. We therefore estimated this using quantities and values calculated from CEIC data, adding up quantities and values for all categories that had quantities, and finding the unit value. We then looked at the growth rate of the CEIC unit values and applied this to CEIC data. The growth rate and unit value for 2004 was consistent with OECD data, providing an overlap in data sets. 12 We calculated “hot money” inflows assuming they started in 2003. Therefore, we find that the difference in 2002 is zero.

5

there is no sign of this at all. Every table shows stable or declining unit prices, and in some cases sharply declining unit prices, for manufactured imports. Eurostat even publishes unit price indexes specific to imports from China; these show no net change for manufacturing or machinery, the major Chinese export sectors. Further, if Chinese exporters had simply raised the prices of goods sold to the world market, we would expect to see a loss of market share. Manufactured goods together total nearly 90 percent of Chinese exports by value, and it seems unlikely that their prices could rise sharply without significant losses in favor of Vietnam, Malaysia and other low-wage competitors. In fact, there were no significant losses toward competing exporters. To the contrary, the quantity measures in manufacturing show robust growth in 2003 through 2006. In the case of commodity-based exports, such as animal and vegetable oils and fats, price increases could have happened without loss of market share, insofar as overall commodity prices rose during this period. But such goods are a trivial share of total Chinese exports by both value and quantity. Table Three shows the ASEAN share of world markets over this period: there is no sign of an increase that could be attributed to Chinese exporters pricing themselves out of the market. Table Three: ASEAN Share in World Exports13 (Percentages)

ASEAN Share in World Exports

1998

1999

2000

2001

2002

2003

2004

2005

4.8%

5.1%

5.6%

5.2%

5.2%

4.0%

5.4%

5.4%

Source: UNCTAD Statistics Database

A second possibility is that China has upgraded the actual quality of its exports, justifying higher unit values not with price increases but with better goods. This possibility is particularly relevant to the processing trade, where China could be importing increasingly high-value goods in order to finish them and export them again. But if this were the case, then unit values of Chinese imports in manufacturing would also be increasing. Table Four gives unit values for imports in manufactures and machinery; no unusual increase is observed, although there appears to be a steady slight progression in unit price from 2001 onward. This may account for part of the export unit price increase, but not too much of it. This share of the unit value increases could be due to an increase in the technological content of process-trade goods, which did occur starting in 2002. However, the fact that the unit value increase is not reflected in Chinese import unit value increases (Table Four) leads us to infer that changes in the composition and degree of the 13

Also according to the UNCTAD Statistics Database, the share in world manufacturing exports for Eastern, Southern, and Southeastern Asia excluding China was 16% in 1995, 20% in 2000 and 23% in 2005.

6

processing trade do not account for the entire unit value increase and perhaps not even for very much of it. Note that the report unit value for manufacturing in 2004 is only 13 percent higher than its 1998 value, while export unit value is up over 50 percent. This would suggest that around three-quarters of the excess increase in export unit value cannot have been contributed by increased unit values in semi-processed imports, even if the processing trade were all of China’s manufactured exports, which it is not.

Table Four: Import Unit Values (into China) (in USD)

Import Unit Values Manufacturing Import Unit Values

1996 0.47

1997 0.38

1998 0.38

1999 0.36

2000 0.36

2001 0.38

2002 0.38

2003 0.43

2004 0.45

0.74

0.52

0.54

0.47

0.43

0.52

0.55

0.59

0.61

Source: OECD Data and authors’ calculations

The next issue is, to what extent did the composition of Chinese exports across broad industrial sectors change? Had the sharp rise in the per-unit value of Chinese exports been the result of a sudden shift in the composition of exports, such shifts would be observed in the data at the level of broad SITC categories. There is in fact an increase of about three percentage points per year in the export share of the machinery and transport equipment sector, but this increase has been going on for a long time, and the gains after 2003 are not out of line with past experience. Otherwise, the sectoral composition of Chinese exports over time appears substantially stable. Processing trade in particular remained steady in percentage of overall exports, at 55% per year from 2000 through 2005, according to the Ministry of Finance14. In the reported data, the surge in dollar value per unit of exports since 2003 originates in the manufacturing sector. Of the total dollar value per unit of all exports in 2004, or $0.66, $0.11 is from the manufactured goods sector, $0.30 is from the machinery and transport equipment sector, and $0.18 is from the miscellaneous manufactured goods sector. Table Five illustrates the patterns of per unit value change across sectors.

14

http://www.mofcom.gov.cn/tongjiziliao/tongjiziliao.html

7

Table Five: Sector Unit Value Representation in Total Dollar Value per Exported Unit, by SITC Code (in US Dollars)

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Total trade 0.51 0.52 0.51 0.47 0.47 0.46 0.46 0.45 0.46 0.56 0.66

0 Food and live animals 0.04 0.04 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02

2 Crude materials, inedible, except fuels 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

3 Mineral fuels, lubricants and related materials 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.02

4 Animal and vegetable oils, fats and waxes 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

5 Chemicals and related 6 products, Manufactured n.e.s. goods 0.03 0.10 0.03 0.11 0.03 0.10 0.03 0.09 0.03 0.08 0.02 0.08 0.02 0.08 0.02 0.07 0.02 0.08 0.03 0.09 0.03 0.11

7 Machinery and transport equipment 0.09 0.11 0.12 0.11 0.13 0.14 0.15 0.16 0.18 0.24 0.30

8 M iscellaneous manufactured articles 0.21 0.19 0.19 0.18 0.18 0.17 0.16 0.15 0.14 0.16 0.18

Source: OECD Data and authors’ calculations

Further, the change in unit export values does not reflect or correspond to any large increase in the wage bill. This can be seen in Table Six, which illustrates wages per unit output. We do not see a marked decrease in this ratio, which would have indicated that a price increase is related to a sudden increase in wage claims. Rather, wage costs appear to hold steady from 1997 onward.

Table Six: Wage to Output Ratio (in Percentages) Year Wage Bill

1996 13%

1997 12%

1998 11%

1999 11%

2000

2001

11%

11%

2002

2003

11%

11%

2004 11%

2005 11%

Source: China Data Center and authors’ calculations

The final issue is, to what degree could shifts in the composition of Chinese exports across narrowly defined (three-digit) industrial categories within manufacturing account for the apparent rise in unit values? Table Twenty-Two in the appendix reports the results of a disaggregation exercise, aimed at isolating those categories with the largest increase in export 8

2006 11%

share due to apparent changes in unit value. We find that compositional shifts do occur, but they do not appear to be very large in relation to the overall increase in reported unit value. And their interpretation is ambiguous. Interestingly, the largest changes are substantially concentrated in capital goods sectors such as machinery and equipment – precisely those sectors where quality changes, quantity increases and disguised capital inflows would be hardest to disentangle. We are left with no clear resolution on this topic, While it is possible that China suddenly and rapidly upgraded the quality of its machinery exports after 2002, it is also possible that those seeking to disguise capital inflow would tend to choose these same sectors as being the safest channel for such activity. Detailed forensic work, case studies of technical change in the relevant sectors, and insider accounts would appear to be necessary to resolve the issue. We infer that China’s export figures overall, as well as in the important manufacturing sectors, are very probably overstated. By how much? We now examine the extent to which funds may have entered China via this illicit route. A clue to the phenomenon at hand may possibly be found in the percentage change in gross capital formation (Table Seven). This figure increases sharply in the post 2002 years. This is the result of an enormous increase in the construction of fixed assets such as plant and equipment, offices and housing. Table Seven: Gross Capital Formation (In Current US Dollars or Percentages where indicated)

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

GDP (billions of current US$) 856.1 952.7 1,019.5 1,083.3 1,198.5 1,324.8 1,453.8 1,641.0 1,931.7 2,243.9 2,668.1

Gross capital formation (billions of current US$) 346.2 361.5 378.2 398.0 420.9 480.5 550.5 676.1 835.7 971.0 1,085.8

9

Percentage change in gross capital formation 4% 5% 5% 6% 14% 15% 23% 24% 16% 12%

Percentage change in gross capital formation adjusted for min hot money inflows

Percentage change in gross capital formation adjusted for max hot money inflows

4% 5% 5% 6% 14% 15% 18% 19% 12% 7%

4% 5% 5% 6% 14% 15% 7% 8% -2% -11%

Source: WDI and authors’ calculations

The increase in gross capital formation reflects, in other words, the construction boom that is everywhere visible in urban China. Gross capital formation increased by more than 60 percentage points from 2003 to 2006. What is more, to take a specific instance, the Beijing real estate industry operating income and profit moves sharply from negative to positive numbers in 2003, a dramatic increase. This is a very significant change. Table Eight: Beijing Real Estate Statistics (in Million Yuan)

2000 2001 2002 2003 2004 2005 2006

Real Estate Industry Operating Income -1862 -1046 -1026 895.9 8661.1 6184.4 11053

Real Estate Industry Total Profits -1303 -215.3 -587.1 1743.3 10701 8131 14959

Investment in Office Buildings 4521.9 7199.3 9732.6 14275 18789 19617 21674

Commercial Buildings Sold 424.84 1245.8 2595.3 5177.9 5883.4 12085 16256

Source: CEIC

The question, then, is: how much of this increase might be accounted for by capital inflow? We believe the answer could be: much of it. If we assume, conservatively, that 30% of the increase in export unit value is due to disguised capital inflows, we find that China is running a much smaller trade surplus. In this case, we estimate that the total disguised capital inflows into the export account were USD $23 billion in 2003, $54 billion in 2004, $95 billion in 2005, and $157 billion in 2006. This accounts for much of the rise in fixed investments as a share of GDP that had occurred up to that point. At the other extreme, based in part on unit price indices reported by importers of Chinese manufactures, it would not be unreasonable to argue that there was no increase in real unit export values after 2002. The disguised capital inflows would amount to $87 billion in 2003, $199 billion in 2004, $347 billion in 2005 and $529 billion in 2006. In that case, China’s 2006 “actual” current account deficit would amount to $425 billion, and the cumulative deficit since 2003, disguised by capital inflow, would amount to $847 billion.

10

Table Nine: Capital inflows under varying assumptions of unit value increase (in USD) Percentage of Increase from 2002

2003

2004

2005

2006

30%

26.0

59.6

104.0.

158.6

60%

52.0

119.2

208.0

317.2

90%

77.9

178.9

312.1

475.8

100%

86.6

198.7

346.7

528.6

Source: OECD, WDI and CEIC Data and authors’ calculations

Table Ten: Effect of estimated capital inflows disguised as current account on Balance of Trade (in Current billions of US Dollars) Balance of Trade Adjusted for Maximum Value (100%) of Hot Money Inflows

Exports of Goods and Services

Exports of Goods and Services Minus Hot Money Inflows

Difference (Max Hot Money Inflows)

Imports of Goods and Services

Balance of Trade

Balance of Trade Adjusted for Minimum Value (30%) of Hot Money Inflows

1998

207

207

0

164

44

44

44

1999

221

221

0

190,

31

31

31

2000

280

280

0

251

29

29

29

2001

299

299

0

271

28

28

28

2002

365

365

0

328

37

37

37

2003

485

398

87

449

36

10

-51

2004

656

457

199

607

49

-10

-149

2005

837

490

347

712

125

21

-222

2006

981

452

529

878

103

-55

-425

Source: WDI and authors’ calculations

11

Caveats and Qualifications The assumptions used in our calculations are subject to several qualifications and reservations. First, we assumed that growth in unit value data from the CEIC database is transferable to OECD data. Since SITC category and overall quantity totals are not given in the CEIC database (only subcategory values are available), the translation may not be entirely accurate for 2005-6. When we asked the OECD statistics division why sub-category totals were smaller than SITC category and overall totals, they replied that this was due to “confidential reasons” on the part of the Chinese government. We incorporated only values and quantities for CEIC categories that had both types of data available, and believe that this may underestimate the unit value increases. Therefore, on this account, we erred on the side of understating capital inflows. Second, there are issues with Chinese reported statistics as noted in other literature. There are problems with GDP, particularly with the overstatement of GDP for political purposes, as well as problems with trade statistics, due to Hong Kong re-exports16. Although this is problematic in determining exact numbers, the phenomenon of inflated export figures is more or less traceable since the last major shift in statistics occurred in 1998, when the National Bureau of Statistics began to use sample survey estimates of small scale industry, affecting the calculation of GDP (Naughton 2007, p. 141). A smaller shift in statistical classification has occurred in the past few years, when several export categories were broken into sub-categories, while some were discontinued. However, this did not affect trade statistics within the larger SITC categories. . Changes in the Financial Environment Assuming that China has, in fact, experienced major capital inflow disguised as export earnings, why did it happen? In part, we believe, because changes in China’s regulatory environment made it possible. In 2003, there were several changes in China’s financial sector which made the environment more favorable to capital inflows. The interest rate began to look more attractive vis-à-vis the dollar, while the NDF premium began to decrease, indicating expectations of yuan appreciation against the dollar. Tables Eleven and Twelve illustrate these trends.

Green writes that the US exaggerates value-added in Hong Kong as around 25% of China’s goods value, while China tends to understate these values. He believes the US-China deficit may be the average of the two records. In any case, China’s understatement of Hong Kong reexports has not changed over time, so does not affect the general unit value trend. 16

12

Table Eleven: RMB Less Dollar Yields (In Basis Points)

Avg 3-month Chinese Repo less US Treasury Avg 3-month CHIBOR less USD LIBOR

1998

1999

2000

2001

2002

2003

2004

2005

2006

1.96

-1.17

-3.4

-0.83

0.54

1.59

1.35

-1.44

-2.41

2.23

0.95

-2.46

0.03

1.6

1.66

1.71

-0.77

-2.57

Source: CEIC, US Treasury Statistics, British Bankers Association

Table Twelve: Non-Deliverable Forward Premium (Percentage of Spot)

Source: Ma and McCauley (2007, p. 16)

13

The rise in unit export values also occurred in conjunction with an important change in the rules governing the holding of dollars inside China. In October 2002, the central government gave permission for all companies to hold foreign exchange accounts. Controls over foreign exchange purchases were relaxed for many businesses, including exporters, while the ability to open foreign exchange accounts was extended to firms outside bonded zones (Lehmanbrown.com, 2002). The goal of this measure was to liberalize the current account, facilitating trade and reducing the state presence in credit markets. Not surprisingly, we see, in Table Thirteen, that foreign exchange transactions within China increased tremendously beginning in 2003. Table Thirteen: Foreign Exchange Transactions within China (in 100 Million Units) Overall Turnover (in USD)

2001 2002 2003 2004

750.3 971.9 1511.3 2090.4

USD Trading Volume

HKD Trading Volume

741.3 951.1 1478.2 2044.1

30.6 108.8 186.3 244.9

JPY Trading Volume

613.9 730.8 761.6 1349.6

EURO Trading Volume

N/A 1.1 3.0 1.9

Source: People’s Bank of China Thus, the regulatory and investment environment was ripe for injecting capital inflows into China. Using the trade account to bring in capital was relatively simple over this period. Exporting companies simply had to overbill exports, and foreign exchange could be transferred into the companies’ bank accounts. The recent crackdown on short-term foreign exchange accounts, and the punishment of both foreign and domestic banks for the violation of exchanging currency outside of controls, has revealed how loose controls over foreign exchange accounts had become. Further evidence comes from the recent exposure and punishment of a large underground bank headquartered in Shenzhen, which exchanged foreign currency and maintained foreign exchange accounts. All of these measures are attempts by the central government to curb hot money inflows and illegal foreign exchange transactions, in order to maintain better control over the current account17. Based on CEIC data available thus far this year, the unit value of exports has virtually ended its upward movement, and perhaps the disguised inflow of capital has now come to an end. Part of the increase, too, may stem from over-billing exports to receive additional Value-Added Tax 18(VAT) rebates after the January 2002 legislation loosened restrictions over VAT rebates. However, in our calculation, we do not see a large unit price increase for the year 2002, which would indicate that VAT abuses due to the legislation have not been very large. 17

In addition, the real appreciation of the renminbi in terms of the dollar in December 2006 signals a change in the desirability of purchasing RMB with dollars. 18 VAT rates range from 5-17%. The standard VAT rate is 17%.

14

What is more, accession to the World Trade Organization affected the trade climate after 2001, but the process of trade and capital control reform continued to be gradual. We believe, then, that much of the export unit price increase is due to overstated values that hide hot money inflows into China’s real estate and other asset markets. Implications of capital inflow via China’s trade account It appears that China ran a true current account surplus much smaller than reported, and may have actually run a trade deficit from 2003 to 2006. This conclusion depends in part on an assumption that China’s import statistics are accurate. Though in the past the over-invoicing of imports may have served to mask capital outflow from China, we have seen no evidence to suggest similar distortions in China’s import accounts at the present time. Indeed there would be little reason for it: hot money flows where markets are hot, as they unquestionably have been in China, and where a currency is widely expected to appreciate. This too was the Chinese case. There are several implications of large net capital inflows disguised as exports. At a glance, trade statistics, and any calculations which use export or net export values, will require correction19. This includes both GDP and the growth rate, figures envied by most other developed and developing nations alike. Also suspect are the recent large profit increases reported by many Chinese firms, which could be an artifact of laundering exaggerated export earnings. Given that we find a much smaller current account surplus, or even a deficit, repeated calls by the United States for appreciation of the RMB based on evidence of an exploding trade surplus lose force. An appreciation of the RMB would increase the exchange value of the hot money invested in China from 2003 to 2006, and the most interested parties may be speculators – including some with political connections in the United States -- who have engaged in illicit investment in China through fictitious trade. This reduces much of the demand for RMB appreciation to an interest in validating a currency speculation. The nature of China’s external financial balance would also change. As an investment on behalf of foreign interests, capital inflow places foreign claims on domestic assets. Although Chinese domestic savings and investment are high relative to other countries, domestic savings and investment will be seen as much lower than they have seemed, while foreign investment will be seen as much higher, once capital inflows through trade are correctly accounted for, Conclusion We believe that simple macroeconomic evidence points quite strongly to a significant overstatement of China’s exports, masking an equivalent capital inflow. This inflow is potentially large enough to put China’s actual current account into deficit, greatly weakening the case for appreciation of the RMB. It also suggests that other aspects of the widely-held view of recent Chinese economic performance, including the profits boom in Chinese enterprises and the growth rate of the economy overall, should perhaps be re-examined for evidence of the role of capital inflow in distorting both the statistics and the underlying economy. 19

Some studies have shown that China’s GDP statistics are overstated for reasons other than errors in the trade account. The overstatement has taken place because some firm managers exaggerate output.

15

References 1. Anderlini, Jamil. “China Hits Out Over ‘Hot Money,’” Financial Times, June 27, 2007. 2. Chinadaily.com. May 19, 2004, http://www.chinadaily.com.cn/english/doc/200405/19/content_332004.htm 3. Glick, Reuven and Hutchison, Michael. “Capital Controls and Exchange Rate Instability in Developing Economies,” Pacific Basin Working Paper Series, Center for Pacific Basin Monetary and Economic Studies, Economic Research Department, Federal Reserve Bank of San Francisco, 2000. 4. Green, Stephen. On the Ground in Asia, Standard Chartered Bank, Shanghai, April 13, 2006. 5. Green, Stephen. On the Ground in Asia, Standard Chartered Bank, Shanghai, May 17, 2007 6. Lehmanbrown.com, 2002. http://www.lehmanbrown.com/FAQ/FAQ-Forex/2.htm 7. Ma, Guonan and Robert N. McCauley. “Do China’s Capital Controls Still Bind? Implications for Monetary Autonomy and Capital Liberalization,” BIS Working Paper No. 233, www.bis.org. 8. Naughton, Barry. “An Economic Bubble? Chinese Policy Adapts to Rapidly Changing Conditions, ” China Leadership Monitor, No. 9, 2003. 9. Naughton, Barry. The Chinese Economy, MIT Press: Cambridge, MA 2007. 10. People’s Bank of China. “People’s Bank of China Adjusted Foreign Exchange Administration Policy towards Individuals,” January 23, 2007, http://www.pbc.gov.cn/english//detail.asp?col=6400&ID=791. 11. Prasad, Eswar and Wei, Shang-Jin. “The Chinese Approach to Capital Inflows: Patterns and Possible Explanations,” IMF Working Paper, 2005. 12. Setser, Brad and Rosenblatt, Casson. RGE China Reserve Watch, February 2006. 13. Wei, Shang-Jin and Zhang, Zhiwei. “Collateral Damage: Exchange Controls and International Trade,” NBER Working Paper No. 13020, April 2007. 14. World Bank. World Development Indicators Online. 15. Xin, Zhiming. “Inflows of Hot Money to be Curbed,” chinadaily, June 27, 2007, http://www.chinadaily.com.cn/china/2007-06/27/content_903359.htm. 16. Zhang, Ran. “Hot Money Influx is ‘Cooling Down,” July 13, 2007, http://www.chinadaily.com.cn/china/2007-07/13/content_5433984.htm. 17. Zhong, Weike. “Control of Exports,” China Chemical Reporter, June 16, 2007, http://goliath.ecnext.com/coms2/summary_0199-6672809_ITM.

16

Appendix Table Fourteen: Change in Unit Price, Year on Year, by Sector (in Percentages)

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Total trade 3% -3% -7% 1% -2% 0% -3% 3% 21% 18%

0 Food and live animals -17% -2% -17% -4% -9% -8% -5% -4% 8% -6%

2 Crude materials, inedible, except fuels -11% -12% -20% -16% 2% -11% -15% -12% 3% 2%

3 Mineral fuels, lubricants and related materials 10% 6% -10% -26% -17% 32% -2% -16% 18% 14%

4 Animal and vegetable oils, fats and waxes -23% -21% 32% -52% -61% -31% -13% -26% 6% 12%

5 Chemicals and related 6 products, Manufactured n.e.s. goods 22% 17% -7% -16% -12% -7% 1% -5% -8% -6% -9% 0% 1% -6% -4% 1% 15% 17% 18% 28%

7 Machinery and transport equipment 20% 7% -5% 15% 8% 10% 5% 12% 33% 25%

8 Miscellaneous manufactured articles -8% -1% -4% 0% -5% -7% -8% -3% 12% 9%

Source: OECD Data and authors’ calculations

Table Fifteen: Export Unit Values of ASEAN Countries20 (in US Dollars)

Indonesia Philippines Singapore Thailand

1996 109 146 120 127

1997 104 134 112 122

1998 81 105 97 107

1999 65 121 96 102

2000 100 100 100 100

2001 90 84 93 102

2002 96 77 91 97

Source: UN Comtrade Yearbook 2005

20

We show all available values for ASEAN countries.

17

2003 103 79 90 105

2004 120 77 93 118

2005 81 N/A 96 130

Table Sixteen: U.S Import Price Indices (from World21) by Commodity Category

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Food and Beverages live and animals tobacco 96 86 93 87 95 87 111 88 105 91 103 93 108 97 103 98 103 100 100 101 95 103 99 103 101 104 112 107 117 109

Crude materials, inedible, except fuels 82 86 91 102 111 106 103 92 101 97 90 95 108 126 134

Mineral fuels, lubricants and related materials 63 60 48 56 59 80 61 38 83 106 61 95 108 141 202

Chemicals and related Manuproducts, factured n.e.s. goods 96 91 97 90 96 90 104 98 106 104 105 98 101 99 97 94 98 97 101 100 97 92 98 94 101 98 110 110 115 114

Machinery and transport equipment 105 106 108 110 112 110 105 102 100 99 98 96 95 95 94

Commodities and Misc. tranmanufactured sactions, articles n.e.s. 98 100 122 101 136 102 137 104 138 103 133 103 110 101 105 101 104 100 95 99 99 99 114 100 139 101 157 101 171

Source: BLS Data22

21

Statistics on US imports from China exist (they are collected by BLS) but begin only in 2004, which is insufficient for our purposes, but even then they show that since 2004, the import price index from China to the US has held steady or is slightly declining. 22 Category 4 was not available, and Category 971 out of 9 was the only available category. Also, some months were missing, so we used data from month 12.

18

Table Seventeen: EU15 Unit Value Index (2000 = 100), Imports from China (excl HK) Food and live animals chiefly Beverag for es and food tobacco 88 67 92 75 97 94 96 97 96 96 100 100 100 100 98 92 87 83 82 75 85 75 94 77

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Crude materials, inedible except fuels 81 85 93 94 90 100 103 91 84 90 97 97

Mineral fuels, lubricants and related materials 87 96 103 102 92 100 127 112 127 261 217 187

Animal and vegetable oils, fats and waxes 82 99 119 125 105 100 98 83 78 84 94 90

Chemical s and related Manuproducts, factured n.e.s. goods 99 87 101 89 103 95 98 92 94 89 100 100 99 101 91 94 82 84 78 83 82 85 83 89

Source: Eurostat Online and authors’ calculations23

23

Yearly data was calculated by averaging monthly data

19

Machinery and transport equipment 92 92 95 91 90 100 100 97 88 88 88 89

Misc. manuTotal factured All articles products 77 83 81 86 89 93 88 90 88 89 100 100 99 100 93 95 82 85 78 84 79 84 85 88

Table Eighteen: Japanese Imports from World, Unit Price Index on Yen Basis, (2000 average=100)

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Metals All Foodstuffs & com& related modities feedstuffs Textiles products 118 122 116 125 111 120 110 114 100 109 98 93 94 107 99 95 94 106 98 100 103 119 106 102 111 123 113 112 105 123 118 108 96 108 107 94 100 100 100 100 102 110 103 101 101 113 103 100 100 116 100 102 104 124 99 125 118 127 100 153 137 135 105 216

Wood, lumber & Petroleum, Chemicals Machinery related coal & & related & Other products natural gas products equipment goods 110 99 103 143 116 115 88 94 139 110 141 73 90 125 100 125 62 95 117 97 117 63 104 110 100 131 80 102 115 107 133 92 111 119 113 111 73 108 122 117 107 71 98 108 105 100 100 100 100 100 104 107 105 97 106 107 105 105 93 107 104 112 110 85 103 111 124 115 80 104 113 172 124 78 109 132 216 139 81 120

Source: Statistics Bureau, Ministry of Internal Affairs, Japan and authors’ calculations24

24

Yearly data was calculated by averaging monthly data

20

Table Nineteen: Change in Unit Quantity, Year on Year, by Sector (in Percentages)

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Total trade 19% 5% 30% 0% 9% 27% 10% 19% 11% 14%

0 Food and live animals -56% 8% 65% 3% -2% 45% -16% 40% 29% -42%

2 Crude materials, inedible, except fuels 0% -7% 17% -9% -7% 30% 35% 46% -31% -1%

3 Mineral fuels, lubricants and related materials 22% -1% 18% -3% -1% 35% 35% -5% 10% -10%

4 Animal and vegetable oils, fats and waxes -20% -7% 71% -58% -63% 9% 15% -19% -29% 15%

5 Chemicals and related 6 products, Manufactured n.e.s. goods 26% 79% 0% -15% 19% 17% -4% -22% 16% -9% 19% 32% 16% -1% 10% 14% 32% 19% 22% 37%

Source: OECD Data and authors’ calculations

21

7 Machinery and transport equipment 25% 13% 26% 17% 29% 31% 2% 30% 18% 32%

8 Miscellaneous manufactured articles 13% 53% 70% 2% 8% 11% 2% 17% 16% 9%

Table Twenty: Change in Total Reported Value by Sector (in Percentages)

1996 2%

1997 21%

1998 1%

1999 6%

2000 28%

2001 7%

2002 22%

2003 35%

2004 35%

0 Food and live animals

3%

8%

-4%

-1%

17%

4%

14%

20%

8%

1 Beverages and tobacco

-2%

-22%

-7%

-21%

-3%

17%

13%

4%

19%

Total trade

2 Crude materials, inedible, except fuels 3 Mineral fuels, lubricants and related materials

-7%

4%

-16%

11%

14%

-6%

6%

14%

16%

11%

18%

-26%

-10%

69%

7%

0%

32%

30%

4 Animal and vegetable oils, fats and waxes

-17%

72%

-53%

-57%

-12%

-4%

-12%

18%

29%

5 Chemicals and related products, n.e.s.

-2%

15%

1%

1%

17%

10%

15%

28%

35%

6 Manufactured goods

-12%

21%

-6%

2%

28%

3%

21%

30%

46%

7 Machinery and transport equipment

12%

24%

15%

17%

40%

15%

34%

48%

43%

8 Misc manufactured articles

4%

25%

0%

3%

19%

1%

16%

25%

24%

9 Commodities and transactions, n.e.s.

-46%

93%

-98%

3022%

154%

18%

15%

49%

-14%

Source: OECD Data and authors’ calculations 22

Table Twenty-One: Share in Value of Total Trade, by Sector (in Percentages) 1996 7%

1997 6%

1998 6%

1999 5%

2000 5%

2001 5%

2002 4%

2003 4%

2004 3%

1%

1%

1%

0%

0%

0%

0%

0%

0%

3%

2%

2%

2%

2%

2%

1%

1%

1%

4%

4%

3%

2%

3%

3%

3%

3%

2%

4 Animal and vegetable oils, fats and waxes

0%

0%

0%

0%

0%

0%

0%

0%

0%

5 Chemicals and related products, n.e.s. 6 Manufactured goods

6% 19%

6% 19%

6% 18%

5% 17%

5% 17%

5% 16%

5% 16%

4% 16%

4% 17%

7 Machinery and transport equipment

23%

24%

27%

30%

33%

36%

39%

43%

45%

8 Miscellaneous manufactured articles

37%

38%

38%

37%

35%

33%

31%

29%

26%

9 Commodities and transactions, n.e.s.

0%

0%

0%

0%

0%

0%

0%

0%

0%

0 Food and live animals 1 Beverages and tobacco 2 Crude materials, inedible, except fuels 3 Mineral fuels, lubricants and related materials

Source: OECD Data and authors’ calculations

23

Table Twenty-Two: Top Twenty Subcategories (within SITC 6 through 8) for Change in Share due to Value Change25 SITC 6 through 8, 3-digit level category 752 Automatic data processing machines, n.e.s. 764 Telecommunication equipment, n.e.s.; & parts, n.e.s. 759 Parts, accessories for machines of groups 751, 752 845 Articles of apparel, of textile fabrics, n.e.s. 894 Baby carriages, toys, games & sporting goods 776 Cathode valves & tubes; diodes; integrated circuits 763 Sound recorders or reproducers; television record. 851 Footwear 778 Electrical machinery & apparatus, n.e.s. 842 Women's clothing, of textile fabrics 821 Furniture & parts; bedding & similar stuffed furni. 841 Men's clothing of textile fabrics, not knitted 893 Articles, n.e.s., of plastics 772 Apparatus for electrical circuits; board, panels 658 Made-up articles, of textile materials, n.e.s. 699 Manufactures of base metal, n.e.s. 848 Articles of apparel, clothing access., excluding textile 653 Fabrics, woven, of man-made fabrics 871 Optical instruments & apparatus, n.e.s. 771 Electric power machinery, and parts thereof

Change in Share due to Value Change, 2002-4 11% 8% 5% 3% 3% 3% 3% 3% 3% 2% 2% 2% 2% 2% 1% 1% 1% 1% 1% 1%

Source: OECD Data and authors’ calculations

To calculate this, we look at the change in the unit value, relative to the average, from 2002-4 relative to the original percentage of total value in 2002. This gave us a percentage that presents the unit-value change component of the shift in share toward the sector. 25

24

The Beijing Bubble: Inequality, Trade and Capital Inflow into China James K. Galbraith Professor Lyndon B. Johnson School of Public Affairs The University of Texas at Austin Austin, Texas Sara Hsu Adjunct Professor of Economics St. Edwards University Austin, Texas Wenjie Zhang LBJ School of Public Affaits The University of Texas at Austin Austin, Texas May 31, 2008 The University of Texas Inequality Project UTIP Working Paper 50 Abstract This paper explores the relationships between inequality, trade and capital flows into China since the early 1990s. We show that the rise in inequality in China since 2000 has more to do with the speculative activities associated with China’s building boom, notably in Beijing, than with the massive growth in manufacturing employment and in Chinese exports since China joined the WTO in 2001. The paper also reports further research on the likelihood of large speculative inflows of capital into China via the current account. An earlier argument for this phenomenon based on inspection of apparent export unit values by sector did not withstand scrutiny in more detailed data sets. Rather, it is the flow of profits from the export boom that has, most likely, fed the speculative fires in the capital and elsewhere.

1

This paper explores the relationships between inequality, trade and capital flows into China since the early 1990s and particularly in the first years of the present decade. We show that the rise in inequality has more to do with the speculative activities associated with China’s building boom, notably in Beijing, than with the massive growth in manufacturing employment and in Chinese exports since China joined the WTO in 2001. Nevertheless, it is the flow of profits from the export boom that has, most likely, fed the speculative fires in the capital and elsewhere.

By all accounts, inequality rose rapidly in China beginning in the early 1990s (Riskin et al., 2001). Measurements by Galbraith, Krytynskaia and Wang (2004) showed that much of the rise in that decade could be attributed to the relative gains of just one province and two municipalities: Guangdong, Shanghai and Beijing, and to the relative earnings gains of just three sectors: transportation, utilities and banking. Major regional losers in relative terms included the Northeast (Manchuria) and the Southwest (Sichuan); across sectors the major losers included manufacturing, farming and trade.i

Figure One presents a broad overview of the evolution of pay inequality in China, overall and by region and sector, through 2005. The method consists of calculating the contribution of each sector within each province to the between-groups component of a Theil T statistic for the whole country, and then aggregating the components by sectors and by provinces to achieve measures of inequality between and within provinces. The figure shows that while during the 1990s inequality between provinces and inequality within provinces (that is, between sectors) both rose, in the 2000s the behavior of these two dimensions of inequality has diverged. Inequality

2

between provinces peaked early in the decade, and has actually declined since 2001. In contrast, inequality within provinces continued to rise.

Figure 1. Inequality between and within provinces in China, 1987-2006.

0.09 0.08 0.07

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

0.06 0.05 0.04 0.03 0.02 0.01 0

1987

Between-Groups Component of Theil's T Statistics

The Overall Theil Inequality Index for China from 1987 to 2006

Year Overall Inequality

Between Provinces

Within Provinces

Source: China Statistical Yearbook and authors’ calculations

Figure Two breaks out the changing inter-regional dimensions of Chinese inequality in a stacked bar graph. Each bar represents a year and each segment represents the “contribution” of a province to overall inequality in that year. Each segment reflects both the population weight of the province (measured by observed employment) and the ratio between average provincial income and national average income. Contributions greater than zero indicate provinces with mean incomes above the national average. Contributions below zero indicate provinces with incomes below average. Overall inter-provincial inequality is measured by the sum of all the elements in a given year; however the statistic is so constructed that longer bars represent higher inequality and vice versa. 3

Figure 2. Contribution of provinces to inter-provincial inequality in China, 1987-2006.

0.20000

Theil Elements between Provinces

0.15000

0.10000

0.05000

0.00000

-0.05000

-0.10000

-0.15000 1989

1988

1987

1991

1990

1992

1993

1994

1996

1995

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Year Beijing

Shanghai

Guangdong

Zhejiang

Jiangsu

Tianjin

Tibet

Qinghai

Ningxia

Hainan

Chongqing

Inner Mongolia

Xinjiang

Liaoning

Yunnan

Gansu

Guizhou

Fujian

Guangxi

Anhui

Shanxi

Jilin

Shaanxi

Hunan

Jiangxi

Sichuan

Shandong

Heilongjiang

Hebei

Hubei

Henan

Source: China Statistical Yearbook and authors’ calculations

The figure shows that the enormous relative contribution of Guangdong province to overall inequality in China actually peaked as far back as 1994, while that of Shanghai reached its zenith around 2000 or 2001. Despite their respective positions as the seat of Chinese export trade and the financial center, both were regressing moderately toward mean income by 2005 -- as incomes elsewhere rose. Uniquely among the big three, the relative contribution of Beijing continued to rise, reflecting in part, no doubt, the acceleration of a program of urban reconstruction and a speculative building boom in advance of the 2008 Olympics. The recent rise of a fourth contender – Zhejiang province – rounds out the contrasting picture of convergence and 4

divergence as the great Chinese coastal development boom matures. able One presents some evidence on trends in manufacturing employment across China during the early years of the new millennium. The table shows that in most Chinese provinces manufacturing employment declined from 2002 through 2006. But there are great exceptions: Guangdong, Zhejiang, Fujian, Jiangsu and Shandong, where manufacturing employment rose by a cumulative total of 4.9 million jobs during these four years. All are deeply involved in China’integration into world markets following accession to the WTO. Their expansion offset a net decline in manufacturing employment of 1.5 million jobs spread across the rest of the country, giving China as a whole a net gain in manufacturing employment exceeding ten percent in that period. Or, in four years these five Chinese provinces added manufacturing jobs equal to thirty-six percent of the remaining manufacturing employment in the United States as of April, 2008.

Table 1. The total number of manufacturing workers by provinces (10000 persons) Source: State Statistical Yearbook The total number of manufacturing workers in the major export provinces (10000 persons) Year/Region Jiangsu Fujian Shandong Guangdong Zhejiang Rest of country 2002 216 134 272 255 97 1933 2003 2004

217 223

152 181

270 280

282 315

109 152

2005 2006

245 281

198 215

334 342

357 387

201 240

1869 1810 1762 1786

It is obvious that these gains in manufacturing employment are closely tied to exports. After rising at just over 10 percent per year, on average, from 1999 through 2001 (two years of boom and one of recession in the US), China’s exports started to surge in 2002. They rose 21 percent that year, and then 35 percent in each of the two following years, before settling back to a reported rate of 28 percent in 2005 and 27 percent in 2006. Overall the reported increase in 5

exports in dollar terms from 2002 to 2006 amounts to a staggering 264 percent. There is no question that a large part of this is “real,” in the sense that reported quantities surged, alongside manufacturing employment in the key exporting provinces.

It is interesting that, apart from the rise of Zhejiang, the post-2001 export boom in China had little effect on inequality as measured between provinces. The explanation is however not far to seek. Though manufacturing in China is a low-wage sector, in high-wage provinces pay in manufacturing can be close to, or even slightly above, national average pay rates. Thus an increase in the manufacturing share of employment would not necessarily increase overall inequality in China: the contribution of a sector whose average pay is close to the national average to overall inequality is necessarily small. This is sufficient to explain why strong growth in export-oriented manufacturing employment need not have had a dramatic impact – one way or the other -- on the inequalities of Chinese society. In contrast, the much higher incomes in banking, utilities, government and real estate in Beijing have a powerful effect on inequality; there is little else in the country quite like them.

Table Two presents the Chinese current account, as officially reported. It may be considered in light of one of the most basic principles of international macroeconomics, that the growth of imports depends on the domestic growth rate, while that of exports depends on growth in external markets. Thus when a developing country experiences a prolonged period of high internal growth, it is normal for a trade deficit to emerge. This is especially likely if the country in question is an importer of food and fuel, and if commodity prices are rising. Innumerable cases can be cited; exceptions, per contra, are rare, and in the modern record largely confined to

6

countries that maintain rigorously undervalued exchange rates and repressed domestic consumption, while rapidly improving the composition and quality of their exports.

Table 2. China’s Balance of Trade, 1998-2006. (Billions of USD)

1998 1999 2000 2001 2002 2003 2004 2005 2006

Exports of Imports of Goods Goods 184 140 195 166 250 225 266 244 323 295 438 413 593 561 762 660 969 791 Source: China Customs

Balance of Trade 44 29 25 22 28 25 32 102 178

Seen from this perspective, Table Two is astounding. China has been running reported internal real growth rates of eight percent or so for three decades, during which time OECD growth rates averaged less than half of that figure (WDI Online). And while during most of this period China reported small trade surpluses, in the most recent years China’s exports have exploded. China’s imports also rose sharply during this period, but exports measured in dollars grew even more, nearly quadrupling from 2000 to 2006: a rise of nearly three-quarters of a trillion dollars. Thus China reported a trade surplus of $103 billion in 2006 in goods and services taken together; the figure for goods alone was $178 billion. Given a dollar value of Chinese GDP at the prevailing exchange rate on the order of three trillion dollars in 2006, exports amounted to nearly a third of GDP by that time and trade openness (exports plus imports) to over half.ii

7

We have examined a number of explanations for this extraordinary turn of events. First, we find no trace of any transformation in unit prices of Chinese exports. Information on the unit prices of imports from China are maintained by European authorities, while the U.S. reports price indices of imports in general. Nothing of consequence seems to have happened in either dataset; indeed dollar prices of Chinese manufactures imported into Europe fell (not surprisingly, given the rise of the euro against the dollar).

Although data we analyzed in a previous working paper (Galbraith, Hsu and Li, 2007) suggested that there was a large (and suspicious) increase in reported unit values of Chinese exports after 2002, further research has deflated this conclusion. Our original hypothesis was that quantity units reported by major product category in OECD summary data could be assumed to be reasonably consistent over short time periods, permitting us to use aggregated, heterogeneous quantities as a rough index of actual shipments. Inspection of the underlying data tables from Comtrade reveals that large changes in reported units did occur (in some instances shifting from actual units to thousands of the same units); thus in most (though not all) categories the hypothesis of extraordinary changes in unit value cannot be sustained.

There have been modest shifts in the composition of exports toward higher-valued goods, notably an increase of about three percentage points per year in the export share of the machinery and transport equipment sector. But this increase had been going on for a long time, and the gains after 2002 are not out of line with past experience. So while China is always in the process of upgrading its manufactured exports, the major push behind the post 2001 boom has been

8

expansion in reported shipments (quantities) rather than in the value-added associated with particular units (price or quality change).

A fourth possibility appears to have better traction: the decline of the dollar. China exports to both Europe and the US, and while the relative stability of the dollar/RMB exchange rate assures that the dollar value of Chinese exports to the US does not fluctuate with the dollar itself, this is not true for Chinese exports elsewhere. Notably, if prices in the final markets do not change, then the euro’s rise would automatically generate larger per-unit dollar earnings for China. But the same effect would work on imports from outside the dollar zone, so it is difficult to see how this artifact of the reference currency would strongly affect the rise in China’s trade surplus.v

A fifth possibility concerns the processing trade, a large share of China’s manufactured exports. China could be importing increasingly high-value goods (from, say, Japan) in order to finish them and export them again. But if this were the case, then reported unit values of Chinese imports in manufacturing would also be increasing, and so would the share of the processing trade in total trade. Neither of these things appears to have occurred. Although there is a slight progression in import prices in manufacturing from 2001 onward, no dramatic increase is observed. Moreover, processing trade accounts for about 55 percent of Chinese exports, and that figure remained stable after 2001.vi If the rise in total export values were due mainly to rising unit value of processed goods, the share of the processing trade in total exports should have risen.

From the remarkable boom in manufacturing employment, coupled to the increase in unit sales, it appears plain that after 2001 China’s exporters took full advantage of their position as a WTO-

9

compliant country, and greatly multiplied their efforts and their results. This calls to mind a comment in Galbraith (2006), following a discussion of the difficulties of making a profit on the manufacture of wage goods for the home market, given the hyper-competitive climate for light industry in China:

“Is there any way for the Chinese manufacturing firm to turn a profit? Yes: the obvious alternative to selling on the domestic market is to export. And export prices, even those paid at wholesale must be many times those obtained at home.”

It would not be surprising, therefore, if an export boom should lead to a profits boom, followed by the speculative concentration of profit incomes in, for example, Beijing real estate. This would appear to be the fundamental mechanism of rising inequality in China in the post-WTO environment.

But if something can be done, it can also be overdone. Since China still maintains capital controls, perhaps Chinese exporters have been over-reporting exports to the Chinese authorities, for the purpose of bringing foreign capital into the country? Perhaps they have been overinvoicing the exports they actually made? Or perhaps they have been, even more simply, reporting exports to the authorities that were never made at all? The next section of this paper considers this possibility, which has been discussed at least to some extent by Chinese officials.

There are straightforward reasons why it would be in the interest of Chinese firms to behave this way, if they could get away with it. The incentive stems from China’s property and stock market

10

booms, and from two regulatory facts: the continued enforcement of controls over capital inflows per se in China, and the legalization, in late 2002, of unlimited foreign currency accounts held in China by Chinese firms. The simple solution from the firm’s point of view in this situation would be regulatory arbitrage: to launder the capital inflow through the current account.

At this point, we are unable to present estimates of the extent to which disguised capital inflow may be occurring; as noted the preliminary estimates in earlier work already cited did not withstand further scrutiny. Detailed forensic work, case studies of the relevant sectors, shipping data, measures of unit imports from the advanced countries, and insider accounts would now appear to be necessary to establish whether regulatory arbitrage is really a large issue. We therefore restrict ourselves, for now, to examining the enabling conditions.

Changes in the Financial Environment

In 2003, there were several changes in China’s financial sector which made the environment more favorable to capital inflows. The interest rate began to look more attractive vis-à-vis the dollar, while the NDF premium began to decrease, indicating expectations of RMB appreciation against the dollar (Ma and McCauley (2007, p. 16). Table Three illustrates the interest rate trends.

11

Table 3: RMB Less Dollar Yields (Percent) Average 3-month Average 3-month Chinese Repo less US CHIBOR less USD Treasury Yield LIBOR 1998 1.96 2.23 1999 -1.17 0.95 2000 -3.4 -2.46 2001 -0.83 0.03 2002 0.54 1.6 2003 1.59 1.66 2004 1.35 1.71 2005 -1.44 -0.77 2006 -2.41 -2.57 Source: CEIC, US Treasury Statistics, British Bankers Association

Further, in October 2002, the central government gave permission for all companies to hold foreign exchange accounts. Controls over foreign exchange purchases were relaxed for many businesses, including exporters, while the ability to open foreign exchange accounts was extended to firms outside bonded zones (Lehmanbrown.com, 2002). The goal of this measure was to liberalize the current account, facilitating trade and reducing the state presence in credit markets. Not surprisingly, Table Four shows that foreign exchange transactions within China increased tremendously beginning in 2003.

12

Table 4: Foreign Exchange Transactions within China (100 Million Units)

2001 2002 2003 2004

Overall USD HKD JPY Turnover Trading Trading Trading (in USD) Volume Volume Volume 750.3 741.3 30.6 613.9 971.9 951.1 108.8 730.8 1511.3 1478.2 186.3 761.6 2090.4 2044.1 244.9 1349.6 Source: People’s Bank of China

EURO Trading Volume N/A 1.1 3.0 1.9

Thus, the regulatory and investment environment was ripe for injecting capital inflows into China. Exporting companies with a willing partner simply had to overstate or over-bill exports, and foreign exchange could be transferred into their bank accounts, from which it could be converted into RMB and used in domestic capital markets.

Did they do so? The recent crackdown on short-term foreign exchange accounts, and the punishment of both foreign and domestic banks for the violation of exchanging currency outside of controls, has revealed how loose controls over foreign exchange accounts had become. Further evidence comes from the recent exposure and punishment of a large underground bank headquartered in Shenzhen, which exchanged foreign currency and maintained foreign exchange accounts. All of these measures are attempts by the central government to curb hot money inflows and illegal foreign exchange transactions, in order to maintain better control over the current accountviii.

Part of the flow, too, may stem from over-billing exports to receive additional Value-Added Tax ix

(VAT) rebates after the January 2002 legislation loosened restrictions over VAT rebates. 13

However, in our calculation, we do not see a large unit price increase for the year 2002, which would indicate that VAT abuses due to the legislation have not been very large.

Profit and Capital Inflows Into Speculative Sectors

We now examine the extent to which these funds – both licit and otherwise -- may have contributed to China’s building boom and particularly to the “Beijing Bubble.” A clue to the phenomenon at hand may possibly be found in the percentage change in gross capital formation. This figure increases sharply in the post 2002 years, while the share of capital formation in GDP rises by seven percentage points between 2001 and 2004. This is the result of an enormous increase in the construction of fixed assets such as plant and equipment, offices and housing. The increase in gross capital formation reflects the construction boom that is everywhere visible in urban China.

Table Five gives the basic information.

14

Table 5: Gross Capital Formation

GDP (billions of current US$)

Gross capital formation (billions of current US$)

Percentage change in gross capital formation (%)

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

856.1 952.7 1,019.50 1,083.30 1,198.50 1,324.80 1,453.80 1,641.00 1,931.70 2,243.90 2,668.10

346.2 361.5 378.2 398 420.9 480.5 550.5 676.1 835.7 971 1,085.80

4 5 5 6 14 15 23 24 16 12

Share of Gross Capital Formation in GDP (%) 40 38 37 37 35 36 38 41 43 43 41

(Current US Dollars or Percent where indicated. Source: WDI and authors’ calculations)

An inflow of export profits, an increase in the profit share in total income, and any foreign capital would need to show up as reported profits in Chinese industry – not only directly but in the sectors ultimately targeted by investment and speculation. This too we observe. To take a specific instance, the Beijing real estate industry operating income and profit moves sharply from negative to positive numbers in 2003, a dramatic increase. Table Six gives the data.

15

Table 6: Beijing Real Estate Statistics (Million Yuan)

2000 2001 2002 2003 2004 2005 2006

Real Estate Industry Real Estate Industry Investment in Commercial Operating Income Total Profits Office Buildings Buildings Sold -1862 -1303 4521.9 424.84 -1046 -215.3 7199.3 1245.8 -1026 -587.1 9732.6 2595.3 895.9 1743.3 14275 5177.9 8661.1 10701 18789 5883.4 6184.4 8131 19617 12085 11053 14959 21674 16256 Source: CEIC

Caveats and Qualifications

We note several further qualifications, arising from Chinese economic statistics as noted in other literature. First of all, there are problems with Chinese GDP, particularly with the overstatement of GDP for political purposes, and with the notorious stability of reported Chinese GDP growth rates. There are well-known problems with the trade statistics, due to the treatment of re-exports from Hong Kongx. There are also problems with achieving continuous measures of trade activity over recent years, due to shifts in statistical classifications, e.g. as several export categories were broken into sub-categories, while some were discontinued. However, this did not affect trade statistics within the larger SITC categories.

Conclusion

While the rise in Chinese inequality seems to have slowed in the middle of the first decade of the twenty-first century, a significant force for continued increases remained, associated with the 16

property boom and other speculative activities that concentrated on the national capital, Beijing, during the period immediately before the 2008 Olympics. The most likely mechanism behind the flow of funds into these sectors is a profits boom associated with the extraordinary increase in Chinese exports that followed WTO accession in 2001. There is reason to suspect that some additional speculative flows occurred by the device of laundering capital inflow through the current account, but despite concentrated efforts we have no firm estimates to offer.

17

Appendix

Tables Seven, Eight and Nine present the sectors with the largest proportionate increases in quantities and values exported. Interestingly, the largest quantity increases are in the metals sectors—rolled steel, pipe, aluminum -- followed by certain electronics sectors. We have no way of judging, a priori, whether an eleven-fold increase in flat-rolled steel exports in four years is plausible or not; we only note that in this and other important sectors the numbers appear to be remarkably high.

These sub-sectors show increases that are very large for such a short time frame, though they are not beyond the bounds of possibility.

18

Table 7: Top 15 Categories (SITC 6 through 8) of Changes in Quantities Exported, 2002-6 SITC Category

SITC Description

Change in Quantity, 2002-6

Dollar change in value 2002-6 (in millions)

673

Flat-rolled products of iron or non-alloy steel, not clad, plated or coated

1126%

7220.2

7643

Transmission apparatus for radio-telephony, radio-telegraphy, etc

653%

29707.5

679

Tubes, pipes and hollow profiles, and tube or pipe fittings of iron or steel

521%

6213.7

7722 7764

Printed circuits Electronic integrated circuits and microassemblies

461% 414%

5847 17148.6

7843

Other parts and accessories of the motor vehicles of 722 and 781-3

380%

7008.8

77812 77121 6531

Electric accumulators (storage batteries) Static converters (e.g., rectifiers) Fabrics, woven, of synthetic filament yarn (not pile and chenille fabrics)

340% 272% 268%

3677.2 -14721.9 7937.6

684 8943

Aluminum Video games of a kind used with a television receiver

251% 242%

5144.9 2795.9

747

Taps, cocks, valves, etc; pressure-reducing, thermostatically control valves

229%

4186.5

7725

Electrical apparatus for switching, protecting electrical circuits, for ≤ 1000 V

227%

3543.7

7649

Parts or accessories suitable for use solely or principally with apparatus of 76

217%

22511.2

763

Sound recorders or reproducers; television image and sound recorders

215%

14825.7

Source: UN Comtrade

19

Table 8: Top 15 Categories (SITC 6 through 8) of Changes in Unit Values of Exported Goods, 2002-6 SITC Category

SITC Description

Change in Unit Value, 2002-6

Dollar change in value 2002-6 (in millions)

87193

Other optical devices, appliances and instruments

1201%

12379.2

793

Ships, boats (including hovercraft) and floating structures

841%

6185.2

761 751 752

Television receivers Office machines Automatic data processing machines and units thereof

493% 270% 220%

10563.2 4637.1 72885.1

7641

Electrical apparatus for line telephony or line telegraphy

195%

6701.3

6531

Fabrics, woven, of synthetic filament yarn (not pile and chenille fabrics)

188%

7937.6

673

Flat-rolled products of iron or non-alloy steel, not clad, plated or coated Pig-iron, spiegeIeisen, sponge iron, iron or steel granules and powders Aluminium Equipment for distributing electricity, nes

186%

7220.2

179%

2026.1

171% 168%

5144.9 5378.3

747

Taps, cocks, valves, etc; pressure-reducing, thermostatically control valves

163%

4186.5

7649

Parts or accessories suitable for use solely or principally with apparatus of 76

162%

22511.2

844

Women's or girls' outerwear, of textile fabrics, knitted or crocheted

161%

7399

759

Parts and accessories (not covers, carrying cases, etc) for machines of 751-52

157’\%

20756.9

671 684 773

Source: UN Comtrade

20

Table 9: Top 15 Categories (SITC 6 through 8) of Changes in Volume-Weighted Unit Values of Exported Goods, 2002-6 SITC Category

SITC Description

Change in Volumeweighted Unit Value, 2002-6 6265%

Dollar change in value 2002-6 (in millions)

87193

Other optical devices, appliances and instruments

673

Flat-rolled products of iron or non-alloy steel, not clad, plated or coated

1304%

7220.2

793

Ships, boats (including hovercraft) and floating structures

1190%

6185.2

761 679

Television receivers Tubes, pipes and hollow profiles, and tube or pipe fittings of iron or steel

897% 351%

10563.2 6213.7

752

Automatic data processing machines and units thereof

341%

72885.1

6531

Fabrics, woven, of synthetic filament yarn (not pile and chenille fabrics)

319%

7937.6

751 684 7764

Office machines Aluminium Electronic integrated circuits and micro-assemblies

285% 248% 213%

4637.1 5144.9 17148.6

7641

Electrical apparatus for line telephony or line telegraphy

208%

6701.3

671

Pig-iron, spiegeIeisen, sponge iron, iron or steel granules and powders

206%

2026.1

747

Taps, cocks, valves, etc; pressure-reducing, thermostatically control valves

205%

4186.5

7843

Other parts and accessories of the motor vehicles of 722 and 781-3

204%

7008.8

191%

5378.3

12379.2

Equipment for distributing electricity, nes 773

Source: UN Comtrade

21

References: James K. Galbraith, “The Firm, Exports, the Banks and the Real Wage in China,” A Note, Prepared for the meeting on Institutional Reform in China, Manchester, UK, August 8-9, 2006. James K. Galbraith, Sara Hsu and Jianjun Li, “Is China Really Running a Trade Surplus?”, UTIP Working Paper No. 45, December 30, 2007. James K. Galbraith, Ludmila Krytynskaia and Qifei Wang, "The Experience of Rising Inequality in

Russia and China during the Transition." European Journal of Comparative Economics, Vol 1, No. 1, 2004.

Carl Riskin, Zhao Renwei, Li Shi, editors. China's retreat from equality : income distribution and economic transition, Armonk, N.Y. : M.E. Sharpe, c2001. Lehmanbrown.com, http://www.lehmanbrown.com/FAQ/FAQ-Forex/2.htm (2002) Ma, Guonan and Robert N. McCauley. “Do China’s Capital Controls Still Bind? Implications for Monetary Autonomy and Capital Liberalization,” BIS Working Paper No. 233, www.bis.org. Naughton, Barry. The Chinese Economy, MIT Press: Cambridge, MA 2007. World Bank. World Development Indicators (WDI) Online. About the Authors: James K. Galbraith, LBJ School of Public Affairs, The University of Texas at Austin. Sara Hsu, Trinity University, San Antonio, Texas. Wenjie Zhang, LBJ School of Public Affairs, The University of Texas at Austin.

i

These results are drawn from data on pay and employment in the State Statistical Yearbook. They are consistent with, but considerably more revealing than, surveys which have tended to characterize the growing gap in Chinese incomes as “urban/rural” or “coast/interior.” ii

The IMF’s world economic outlook pegs nominal GDP for China in 2006 at $2 trillion, in comparison to which the official trade statistics look even larger. http://www.econstats.com/weo/index_glweo.htm. We have not tried to unravel the discrepancy, except to say that all such comparisons are clearly open to skeptical appraisal. v

Were the reference currency switched to the euro, the rise in China’s export earnings would appear lower, since the country’s exports to the U.S., measured in euro, would have been sharply cut by the dollar devaluation. Similar effects would apply on the import side: China’s eurozone imports would not have risen so much, while its dollarzone imports would have risen considerably more. vi

http://www.mofcom.gov.cn/tongjiziliao/tongjiziliao.html

22

viii

In addition, the real appreciation of the RMB in terms of the dollar in December 2006 signals a change in the desirability of purchasing RMB with dollars. ix

VAT rates range from 5-17%. The standard VAT rate is 17%.

x

Green writes that the US exaggerates value-added in Hong Kong as around 25% of China’s goods value, while China tends to understate these values. He believes the US-China deficit may be the average of the two records. In any case, China’s understatement of Hong Kong re-exports has not changed over time, so does not affect the general unit value trend.

23