CHINA BY THE NUMBERS: HOW REFORM AFFECTED CHINESE ECONOMIC STATISTICS⁽¹⁾

Thomas G. Rawski Department of Economics

University of Pittsburgh

Pittsburgh PA 15260 USA

email: tgrawski+@pitt.edu

Revised December 20, 2000

INTRODUCTION

At the outset of reform, China, a poor nation with limited development of information resources, possessed a statistical system that provided its government with a wide array of reasonably accurate quantitative information about economic activity within China's enormous land mass.⁽²⁾ This was partly a reflection of a long tradition of literacy and record-keeping, and partly the result of sustained effort on the part of China's socialist state.

China's pre-reform statistical system produced information that, except for the epidemic of false reporting linked with the Great Leap Forward of 1958-60, withstood scrutiny from an often skeptical international community. Nevertheless, China entered the reform period with important limitations on its capacity to produce timely, accurate, and useful measures of the level, composition, and growth of economic activity.

At the start of reform, the primitive state of China's information and communication industries restricted statistical capabilities. Implementation of new technologies linked to telecommunications and computers has rapidly eroded the limitations on statistical capabilities imposed by inadequate "hardware," which receive no further discussion here.

China's statistics and data-gathering institutions were designed as aids to the management of an administratively controlled, non-market economy. Without substantial reform, new developments in the structure and operation of China's economic mechanism threatened to push the statistical system toward obsolescence and irrelevance. As China's open door initiatives blossomed from slogan to reality at an astonishing rate, the gap between Chinese statistical categories, based largely on eccentric Soviet conventions, and standard measures familiar to researchers and policy-makers in rich and poor nations alike imposed growing costs on China's government and society.

Prior to reform, the Chinese state exercised tight control over the dissemination of statistical information. Policy analysis was seen as a "prohibited zone" that remained largely off-limits to Chinese researchers.⁽³⁾ This situation discouraged active dialogue between producers and consumers of statistics. The government's reflexive secrecy limited Chinese as well as overseas economists to patchwork appraisals developed from isolated scraps of information.

Two decades of reform have brought dramatic change. Publication of the first national statistics yearbook in 1982 signaled the demise of economic secrecy, which vanished amidst a deluge of yearbooks, census materials, surveys and, most recently, internet publications. Official agencies still churn out libraries of information classified as "internal use" (neibu) or "top secret" (jimi, juemi), but occasional glimpses confirm that, although sometimes offering more detail than open publications, secret materials mostly replicate or anticipate information that eventually enters the public domain. Leaders of the National Bureau of Statistics Bureau (NBS - formerly the State Statistics Bureau) agree privately that the classified yearbooks contain no secrets.

Despite - or perhaps because of - the immense volume of data now published by the NBS, and by China's ministries, government bureaus, research organizations, provinces, municipalities, counties, and even individual companies, it remains difficult for the international research community to formulate a consistent view of China's economic achievements and prospects. Lack of consensus is particularly evident as this paper is written. While international bodies like the World Bank, the Asian Development Bank and the International Monetary Fund echo official Chinese claims of continued growth at world-leading annual rates of 7-8 percent, segments of the international business press display growing skepticism about the viability of China's high-speed growth.⁽⁴⁾ Chinese economists and journalists, while not openly challenging the official growth estimates, routinely describe recent economic conditions with language suggesting that domestic as well as overseas observers harbor grave doubts about the validity of recent economic statistics.

This essay probes these issues through a trio of case studies in the measurement of economic activity. I focus successively on urban unemployment, industrial production, and the reliability of recent macroeconomic aggregates.

URBAN UNEMPLOYMENT

China is no exception to the rule that unemployment (and also employment) is notoriously difficult to measure in any economy with a large farm sector.⁽⁵⁾ To this day, there is no comprehensive measure of nation-wide joblessness that can claim any degree of plausibility. Yet the development of unemployment statistics reveals a pattern of dynamic response to public demand for information that illustrates the growing capability of China's statistical system.

In the initial stages of reform, Chinese spokesmen and researchers hesitated to acknowledge the existence of unemployment. The first statistics yearbook, however, gave figures for both the number and the origins of newly employed urban workers; the sources of new workers included "urban persons awaiting employment" (chengzhen daiye renyuan).⁽⁶⁾

China's first compendium of labor statistics, published in 1987, showed annual figures of "urban persons awaiting employment" as well as the ratio of such persons to the total of employed and non-employed persons. The definition of this ratio exactly parallels the standard concept of unemployment rate.⁽⁷⁾ The yearbook of labor and wage statistics initiated in 1989 (with 1988 data), continued this practice. The 1990 edition added provincial as well as national "urban rates of awaiting employment."⁽⁸⁾ The 1991 yearbook added an English Table of Contents, which described these data as "Unemployment in Cities and Towns and Unemployment Rate by Region."⁽⁹⁾ The 1994 yearbook concluded the terminological transition by introducing new Chinese terms: "unemployed persons" (shiye renyuan) and "jobless rate" (shiyelue).⁽¹⁰⁾

These changes in terminology signaled a growing openness to objective, empirically-based research and policy analysis related to issues of employment and unemployment. Amid growing concern about the need for urban enterprises, particularly in the state sector, to shed millions of redundant workers, scrutiny of labor statistics intensified. Researchers discovered that the NBS urban employment surveys underestimated the scale of unemployment. The actual rate of unemployment among registered urban residents was 3.5 and 4.0 percent in 1994 and 1995 (vs. published rates of 2.8 and 2.9 percent).⁽¹¹⁾ The reason: the official figure "comes from looking only at people who register as unemployed with the labor departments."⁽¹²⁾

During the second half of the 1990s, massive increases in the scale of joblessness propelled unemployment to the top of China's economic policy agenda. Explosive growth of unemployment arose from the combined impact of declining economic growth and accelerating layoffs of redundant workers, exacerbated from 1997 by the impact of the Asian financial crisis. These changes opened new gaps between official statistics and actual conditions. NBS adjusted its data to reflect new circumstances. Beginning with the figures for 1998, statistics showing the number of "staff and workers" (zhigong) no longer include workers on "furlough" (xiagang) - implicitly recognizing that these workers face eventual dismissal.

Despite these adjustments, dissatisfaction with official data, which still give unrealistically low urban unemployment figures, spurred researchers outside NBS and the Ministry of Labor to create alternative estimates of joblessness. Hu Angang, for example, includes furloughed workers in his estimates of urban unemployment. Hu calculates a 1997 urban unemployment rate of 5.7 - 7.0 percent, well above to the official figure of 3.1 percent.⁽¹³⁾ Even Hu's figures are too low because they exclude unemployment among migrants who do not possess official urban residence permits. Hu also makes no provision for workers idled by plant closures that affected about 7 million workers in 1995⁽¹⁴⁾ and surely increased thereafter. New studies provide further detail and added refinement.⁽¹⁵⁾

Ironically, gradual implementation of the new measures does not seem to have improved the accuracy of China's industrial statistics. China's Statistics Law, passed in 1983 and amended in 1996, forbids false reporting. Enforcement efforts have uncovered large numbers of violations, many involving exaggeration of industrial production. Critics concentrate their fire on rural industry, complaining that "fiddling with industrial output figures has become rampant among township enterprises."⁽¹⁷⁾ China's 1995 industrial census revealed false reporting on a grand scale. Vice-Premier Zou Jiahua announced that "industrial developments. . . are quite different from the data the government had previously been provided. . . . After striking out phoney [sic] statistics, the listed number of township enterprises. . . decreased by one-third. . . and their output figure dropped by nearly 40 per cent."⁽¹⁸⁾ Two years later, the Ministry of Agriculture appears to have made another downward revision to the 1997 output and employment data for township and village industries.

What are the sources of these difficulties? Multiplication of the number and the variety of industrial enterprises, rapid restructuring of the statistical system, including the new focus on value added, repeated changes in the definition of key indicators, and growing demands to execute surveys and collect census information seem to have overwhelmed the capabilities of local and regional statistical agencies.

At the same time, the shift from plan to market has redirected the attention of enterprise leaders away from government mandates, including statistical reporting, toward clients, sales, and profits. Many enterprise leaders "pay little attention" to statistical reporting, which is widely seen as a burden that "does not produce the information that operating units really need."⁽¹⁹⁾ Statistical agencies complain that firms assign often untrained staff to compile statistics, look for chances to cut positions assigned to statisticians, and refuse to submit standard reports.⁽²⁰⁾

Although the compilation of industrial statistics depends on the data collected and processed by enterprises and local governments, the leaders of these entities often value statistics for their political impact rather than their precision. Just as the mayor of Detroit "had enormous distrust" for the United States population census of 1990 and "challenged the count at every point" until "months of wrangling" produced enough revisions "to lift Detroit above the one million mark,"⁽²¹⁾ Chinese business executives and local officials manipulate statistics to attain the prestige and opportunities attached to large output value, high growth, or whatever quantitative indicator captures public or official attention.

Thus in May 1999, local leaders in Shaoyang, Hunan, announced plans to boost the number of firms producing RMB 10 million to over 50, to raise the number of enterprises with output above RMB 50 million to 6, and to create one or two enterprise groups with annual output of RMB 100 million - all during 1999.⁽²²⁾ In August of the same year, Agriculture Minister Chen Yaobang, accompanied by top provincial leaders, conducted a five-day inspection of rural Hunan. The Minister's itinerary included visits to the "Hardware Township for the Southeast" and other sites in Shaoyang, during which "the more he saw, the happier he became." The visit gave local leaders a chance to socialize with provincial dignitaries and with the Minister, who praised the local leadership for implementing policies that "build the county through commerce, jointly promote industry and trade, and use the market to build business."⁽²³⁾ As Shaoyang's leaders basked in the Minister's praise, were they concerned about possible inaccuracies in local output totals? When Detroit's mayor learned that national officials had placed Detroit's population at 1.027 million, was he worried that this might overstate the actual head count?

In contrast to the outcry over the inadequacies of official data on urban unemployment, we see little public response, even among economists, to widely remarked flaws in China's industrial output statistics. The reason is probably that measures of industrial activity, unlike measures of unemployment, have no direct connection to pressing social issues. After the 1995 industrial census revealed large overstatements, NBS appears to have made some back-of-the-envelope adjustments to figures for rural industry during 1991-95. One writer complained that overseas analysts devote more attention than Chinese researchers to inconsistencies within the industrial data: many people doubt the NBS figures, but few examine them closely.⁽²⁴⁾

Under these circumstances, the National Bureau of Statistics evidently opted to reform the collection of industrial statistics rather than try to eliminate what appears to be a persistent tendency toward exaggeration (shuifen) in the traditional system of vertical reporting. The new system is designed to move the collection of industrial data into channels that effectively bypass local and provincial statistical agencies. The new system will rely on a combination of sample surveys for small firms and direct electronic reports from state firms and from large enterprises outside the state sector.

Before the new structure was fully in place, the system for gathering industrial data was faced a new an completely unexpected challenge: the "wind of falsification and embellishment" (jiabao fukuafeng) that rocked China's economy in 1998.⁽²⁵⁾

CHINA'S STATISTICAL DEBACLE OF 1998

In February 1999, China's National Bureau of Statistics (NBS) publicly rejected provincial figures for growth of gross domestic product (GDP) in 1998. Although provincial figures, reproduced in Table 1, implied national growth amounting to nearly ten percent, the Bureau attributed this outcome to "cooked local figures," expressed its determination to "reduce. . . dependence on the calculations of local governments," and announced that it had "squeezed out the over-reported part" to arrive at an "accurate" total of 7.8 percent growth for 1997/98.⁽²⁶⁾

But the National Bureau itself seems to have run afoul of "winds of falsification" that have buffeted China's statistical reporting system ever since Beijing established the 1998 objective of 8 percent annual growth as a "great political responsibility." The revised official figures for 1997/98 bristle with anomalies. Is 7.8 percent GDP growth during 1997/98 consistent with a 1.6% drop in energy use?⁽²⁷⁾ Could retail sales of consumer goods rise faster than household incomes even though researchers find a sharp drop in the marginal propensity to consume?⁽²⁸⁾ Could industrial output grow by 10.75 percent in 1997/98 when 80 of 94 major products fell short of double-digit growth and 53 suffered output declines?⁽²⁹⁾ Similar difficulties surround official claims about the growth of agriculture, investment, transportation, and retail sales.⁽³⁰⁾

Chinese commentators explain what happened:

Some of the targets that come down from the higher levels are objectively impossible to reach, but since the leaders demand high speed, then the operating departments split up the responsibilities, and, in order to ensure the achievement of the result specified by the upper levels, the lower levels apply more pressure. . . . plan indicators that are based on the requirements sent down by the upper levels in reality are forced on the lower level statistical figures and then returned upwards.⁽³¹⁾

Zhang Sai. . . a former head of the State Statistical Bureau is very concerned about administrative interference with statistical work. . . . the challenge of keeping statistics accurate was particularly difficult last year [i.e., in 1998]. . . . if administrative departments involve themselves too much in statistical work, it will be hard to avoid introducing errors into the data.. . . .. 'exaggerating the size of the task as the order works its way down level by level and the exaggeration of economic performance level by level as economic data works its way towards the center'. . . . . Zhang Sai stressed that an end must be put to the phenomenon of 'officials make statistics and statistics make officials' and 'using exaggeration to win an official position'. ⁽³²⁾

Thus Beijing "issued orders to every province and city. . . . the task assigned to Shanghai is to ensure economic growth of 12 per cent. To this end, the Shanghai government also issued quotas to each level . . . the plans that cannot ensure 12 per cent growth must all be returned for amendment."⁽³³⁾

These and other reports indicate that statistical falsification, previously confined to minor fudging, occasional gross distortions by specific enterprises or localities, and, as noted above, widespread exaggeration of output in rural collective industry, suddenly blossomed into a massive nationwide phenomenon in the second half of 1998. The timing could not have been worse, for 1998 was a year in which the State Council's wholesale reorganization of central government offices meant that "units undertaking sectoral statistics have universally experienced cutbacks, including big reductions in the number of statistical personnel, which sharpens the conflict between statistical capabilities and statistical responsibilities."⁽³⁴⁾ The result was "universal falsification of statistics, as a 'statistical bubble' works its way up through the system, and provides mistaken reportage to the decision-making levels."⁽³⁵⁾ Former NBS Director Zhang Sai attacked the 1998 results in the harshest terms: "Deceiving the nation and tricking the people can lead to untold disasters. . . . Didn't the Great Leap Forward of 1958 teach us a hard enough lesson?!"⁽³⁶⁾

Economists use three methods to calculate a nation's total output. In principle, each calculation delivers the same result. Although actual totals do not coincide, multiple approaches provide valuable opportunities for detecting errors. On the output side, total product is the sum of value-added in various sectors of the economy: agriculture, mining, manufacturing, utilities, construction, transport and communication, and various service industries. The expenditure approach tabulates aggregate output as the sum of final demand components: private consumption, investment, government consumption spending, and net exports. The income account derives the same total by summing flows of wages and salaries, payments of interest, rent, and dividends, indirect taxes, and business profits.

Official measures of 1998 economic performance appear to overstate actual growth despite downward adjustments by NBS. The Bureau's calculations, and indeed the entire reporting system, focus on the output and expenditure measures of total product. The tripartite structure of the national accounts permits us to seek an alternative measure on the income side, which receives little attention in Chinese statistical compilations, and may therefore have escaped the political pressures that led to "winds of falsification and embellishment."

Data from the rural household survey provide an opportunity to test the idea that income data may be relatively free of distortion. Rural survey results for 1997 and 1998 show declining per capita net income from household business operation (including farming) in Helongjiang, in all the Yangzi provinces from Hubei on down to Shanghai, and in several other provinces.⁽³⁷⁾ Falling household income is a natural consequence of major flooding of the Songhua and Yangzi rivers. This provides a more plausible picture of the flood-stricken farm economy than the output statistics, which show rising farm production in all but one province during 1997/98 despite floods that one source ranks among China's top ten natural disasters of the 20^th century.⁽³⁸⁾

Encouraged by this outcome, I have used income-side data, mostly from the 1999 edition of China's annual statistics yearbook, to calculate an alternative measure of GDP growth for 1997/98.⁽³⁹⁾ Table 2 summarizes the results, which show nominal growth of 4.6 percent and, after adjusting for price deflation, real growth of 5.7 percent for 1997/98.

Unfortunately, the new figure of 5.7 percent cannot provide a reliable index of China's economic growth during 1997/98. There are two difficulties. First, the calculations underlying Table 2 are not complete. The missing data cover a portion of wage payments and profits in the urban and rural informal sectors. Revision and reorganization of data for rural enterprises also create the possibility that our results include only part of total profits in this important sector. In both cases, knowledgeable observers suggest that the omitted categories may have grown faster than the economy as a whole.⁽⁴⁰⁾

This is certainly possible. We know that private-sector employment increased during 1997/98. It also seems likely that neither the rural nor the urban survey statistics, which are used extensively in the calculations underlying Table 2, attain full coverage of profits from private enterprise. If the urban and rural surveys undercount earnings from informal employment, and if unrecorded profits from private business increased during 1998, a more complete calculation including data for these missing categories would raise estimated GDP growth for 1997/98 above the 5.7 percent figure shown in Table 2.

A second problem arises because the results in Table 2 are derived almost exclusively from official data. Even though the income-side data offer a more plausible picture of 1998 outcomes than the tainted figures for output and expenditure, there is no basis for assuming that the figures used to derive the new GDP estimate completely escaped the political pressures that engulfed China's entire system of economic measurement beginning in 1998.

Consider the example of taxes - the fastest growing segment of the calculations shown in Table 2. Official fiscal data show that indirect tax collections rose by 13.9 percent during 1997/98, a rate far above even the exaggerated provincial growth figures (Table 1).⁽⁴¹⁾ Since, according to the State Administration of Taxation, "in the first half of the year revenue growth basically matched the country's growth," the official data imply that tax collections spurted far ahead of economic activity during the second half of 1998.⁽⁴²⁾

Efforts to increase tax collections, like efforts to attain 8 percent GDP growth, were reinforced with the trappings of political campaigns, complete with escalating targets. On October 7, the goal was to increase annual revenues by RMB 80 billion. Six weeks later, the target had jumped to RMB 100 billion.⁽⁴³⁾ The intensity of the revenue campaign is evident from the statement that "tax bureaus tried hard to maintain a year-on-year increase of 27.6 per cent in both August and September, even though devastating floods hit many parts of the country."⁽⁴⁴⁾ Under these circumstances, falsification of revenue figures becomes a distinct possibility. Knowledgeable individuals confirm that tax figures are subject to exaggeration (shuifen). One informant suggests that as much as one-fourth of the reported increase in revenue collections for 1997/98 could be fictitious; written sources confirm that local governments use borrowed funds to conceal tax shortfalls.⁽⁴⁵⁾ Finance Minister Xiang Huaicheng's observation that "we have maintained a certain revenue growth rate under very difficult conditions" and his comment that "fiscal and taxation work is quite disorganized and loosely supervised" might be construed as alluding to inflation of accomplishments as well as "problems of tax evasion, tax fraud, tax deferment, and tax default."⁽⁴⁶⁾

This discussion shows that the 5.7 percent growth figure for 1997/98 could be subject to downward as well as upward corrections. Recalculation of 1997/98 growth under the assumption that one-fourth of the reported increase in indirect tax collections was spurious, for example, would reduce estimated real growth to 4.9 percent, a drop of nearly one percentage point.

Other downward corrections might be needed. Given the large income gaps between rich and poor rural regions and urban households, survey workers (or their supervisors) can easily manipulate results for such variables as average income by altering the proportion of respondents from prosperous districts or from households with laid-off workers. Can we be confident that the survey data used to derive several components of our income-side estimates remain untouched by the political pressures that have temporarily unhinged much of China's statistical reporting system? In the absence of clear answers to such questions, we see that the results shown in Table 2 could overestimate 1997/98 growth, perhaps by a large margin.

The obvious conclusion is that any estimate of China's GDP growth for 1997/98 is subject to a wide error margin. This caveat surely applies to the present income-side estimate placing 1997/98 growth of China's real GDP at 5.7 percent. My own belief is that actual growth in 1997/98 was considerably less than the 5.7 percent figure shown in Table 2. Consider the following:⁽⁴⁷⁾

"Per capita income in urban and rural areas continued to fall in the first quarter of this year [1999]."

"A report from National Bureau of Statistics indicates the Chinese people are uncertain about their near-future incomes. In October [1999], 66 per cent of consumers said their household incomes had either remained unchanged or had decreased during the previous 12 months."

Chinese airlines engaged in massive price cutting during 1998. Despite big reductions in the absolute and relative price of air tickets, passenger traffic volume (in passenger-km) rose by only 3.4 percent overall and 2.3 percent on domestic routes.

Each of these reports seems flagrantly at odds with the revised income-side outcome of 5.7 percent growth for 1997/98 as well as the adjusted official figures of 7.8% (for 1997/98) and 7.1% (for 1998/99).

Whatever the outcome for 1998, the "wind of falsification and embellishment" continued to wreak havoc with China's economic statistics during 1999. Once again, the National Bureau publicly rejected the provincial figures for annual GDP growth, which it found to be filled with exaggeration.⁽⁴⁸⁾ Once again, preliminary figures suggest that, despite NBS' efforts to "squeeze out the excess from the statistical data" and to "expose overblown claims," the official figures for 1999 are likely to contain "material indicators. . . markedly at variance with the economic growth rate" derived by NBS.⁽⁴⁹⁾

Official figures in the annual communique on national economic development for 1999 indicate declining annual growth for several GDP components: investment (annual growth down sharply); industry (a small decline); grain, cotton, and sugar (modest declines); and the balance of trade (a big decline due mainly to large-scale redirection of imports into legitimate channels). Indications of a reduction in growth are offset by reported increases in annual growth for transport and communication, retail sales, and enterprise profits.⁽⁵⁰⁾ These figures, particularly the steep decline in reported annual growth of investment spending from 13.9 percent in 1997/98 to 5.2 percent in 1998/99 seem inconsistent with real GDP growth of 7.1 percent, as does information from the State Bureau of Domestic Trade indicating that the share of commodity markets suffering from excess supply rose successively from 25.80 percent in the first half of 1998 to 33.77 percent in the second half of 1998, to 72.40 percent and 80 percent in the first and second halves of 1999, and to 100 percent in early 2000.⁽⁵¹⁾

CONCLUSION

Two decades of reform have produced impressive achievements in the realm of economic statistics. Twenty-five years ago, Chinese economic performance was shrouded in mystery and remained impenetrable to all but a tiny coterie of analysts immersed in the arcane world of "China-watching." By the mid-1990s, the range, depth, and quality of statistical documentation surrounding China's economic performance matched or exceeded comparable materials for many nations at similar or even higher levels of economic development. These improvements were not an automatic consequence of China's gradual shift from plan to market. They required careful planning, extensive mobilization of human and material resources, and sustained effort. China's immense size multiples the task of creating or revising systems of data collection. It also multiplies the scale of reform achievements in improving the measurement of economic activity.

Our three case studies illustrate different facets of the reform process. Changes in the measurement of urban unemployment and industrial output represent typical reform episodes. In each instance, government agencies initiate efforts to improve performance -- in this instance by reforming the compilation of important economic indicators. Reform may face opposition from individuals or interest groups who are satisfied with prior arrangements or fear the consequences of change. Forces originating outside the reform process may accelerate the pace of change -- as we see in the case of unemployment data. The history of efforts to improve the measurement of industrial output shows how social forces unrelated to reform can blunt policy initiatives -- in this case because enterprise leaders and local politicians either have little interest in industrial data or value them for purposes other than accurate measurement.

Despite substantial and impressive achievements, our case studies show that the reform of Chinese economic statistics remains a work in progress. The long-term task is to address shortcomings inherent in the concepts, data collection procedures, and analytic methods underlying current statistical indicators. Chinese statisticians and NBS administrators are keenly aware of these difficulties. NBS has demonstrated its capacity to design new systems and to mount the sustained effort needed for implementation.

The short-term task is to overcome the epidemic of false reporting that has engulfed China's statistical apparatus since 1998 and will soon enter its fourth year. It seems unlikely that the National Bureau of Statistics can resolve the falsification problem on its own. The omission of a GDP growth target in Premier Zhu Rongji's March 1999 annual report and the Premier's comment that "falsification and exaggeration are rampant" show that China's top leaders recognize the source and the persistence of recent difficulties with China's economic statistics.⁽⁵²⁾

The government apparently hopes that relaxing its demands for specific growth targets and public exhortations to "build a clean and honest government" can bring this episode to a close without an embarrassing acknowledgment that official data have overstated recent economic achievements. This approach may achieve a gradual return to normal reporting. In the meantime, the uncertainty surrounding official economic data creates new burdens for policy-makers and researchers.

Following the "Great Leap Forward" of 1958-60, China suffered a calamitous famine, in no small part because no journalists and only a handful of leaders dared to contradict official claims of bumper harvests and overflowing warehouses. Amartya Sen, the Nobel laureate, often contrasts China's controlled media and India's free press, plausibly arguing that India's press would never ignore mounting evidence of impending famine or other national disaster.⁽⁵³⁾

Despite its substantial cost, China's current epidemic of statistical falsification has revealed an important achievement of the past two decades of reform. Reflecting the atmosphere of increasing openness epitomized in the slogan "seek truth from facts" (shishi qiushi), Chinese journalists and economists have not ignored the "wind of falsification and embellishment." As early as November 1998, the NBS journal Zhongguo Tongji [China Statistics] carried a scathing attack, cited earlier, on the consequences of imposing unrealistic economic targets. In the same month, Liu Guoguang, a prominent economist and policy adviser, writing in Renmin ribao (People's Daily), said "there is no need to pursue 8 percent to excess. . . . it could encourage a spurious atmosphere by causing the lower levels to send in false reports."

Even as China's leaders repeat dubious claims of continued high economic growth to domestic and international audiences, new voices join the attack on false statistics. Thus Wang Xiaolu writes that "Correct macroeconomic control must be built on accurate economic information. Mistaken judgments about the economy during the past several years are related to unrealistic statistics. Despite immense efforts by the statistical departments, the phenomenon of statistics departing from reality still exists to a certain extent."⁽⁵⁴⁾

No Chinese economist has publicly stated what many believe -- that beginning in 1998, official measures of China's economic output growth intentionally exaggerate actual performance. Officials as well as researchers hesitate to speak out in specific terms. When told of former NBS Director Zhang Sai's attack on false statistics, one highly placed individual said "he is retired, so he can say whatever he likes."

Even so, the public response to this episode reveals a social milieu that is very different from the closed system of the 1960s and 1970s. Our review shows that in 1998, unlike 1958, 1968, or 1978, important social, political, or economic phenomena could no longer lie hidden behind a curtain of official secrecy and academic timidity. Even though China's journalism is a far cry from the rough and tumble of India's, we see that reform has eliminated the distinction that Sen observed. The continuing growth of information networks gives ample reason to expect that this important stride toward an open society will not be reversed.

Table 1

China GDP Growth During 1997/98 and 1998/99 Provincial and National Figures

Sources: growth rates for 1997/98 from Xu Binglan, op. cit; for 1998/99 from "NBS Deals a Blow to Falsification," dated 28 February 2000, retrieved from the web site of Zhongguo qingnianbao [China Youth Daily], www.cyd.com.cn. Provincial GDP totals for 1997 are from Zhongguo tongji nianjian 1998 [China Statistics Yearbook], p. 62. Weighted averages use 1997 provincial GDP figures as weights.

Income-Side Estimate of Chinese Economic Growth During 1997/98

Source: Files: \prc\gdp\gdpcalc.304.cls and \prc\gdp\gdpnotes.304.doc posted at http://www.pitt.edu\~tgrawski

1. The author gratefully acknowledges research support from the American Council of Learned Societies, the National Endowment for Humanities, the University of Pittsburgh's University Center for International Studies, and Harvard University's Fairbank Center for East Asian Research. The author has also benefitted from information, comment, and advice provided by Zheng Jingping, by other participants in the May 2000 conference on "Key Issues in China's Economic Development and the Use of Statistical Data," and by Loren Brandt, David Cowhig, Kenichi Imai, Gary Jefferson, Albert Keidel, Nicholas Lardy, Barry Naughton, Carl Riskin, Kai-yuen Tsui, Harry X. Wu, and a number of China-based colleagues. Several of these colleagues strenuously dispute certain views advanced here, which are the sole responsibility of the author.

2. International acceptance of official Chinese data dates from the publication of Alexander Eckstein ed., Quantitative Measures of China's Economic Output (Ann Arbor: University of Michigan Press, 1980). Earlier discussions of data issues include Choh-ming Li, The Statistical System of Communist China (Berkeley: University of California Press, 1962); Shigeru Ishikawa, National Income and Capital Formation in Mainland China (Tokyo: Institute of Asian Economic Affairs, 1965); Ta-chung Liu and Kung-chia Yeh, The Economy of the Chinese Mainland (Princeton: Princeton University Press, 1965); Dwight H. Perkins, Market Control and Planning in Communist China (Cambridge: Harvard University Press, 1966); Nai-ruenn Chen, Chinese Economic Statistics (Chicago: Aldine, 1967); Thomas G. Rawski, "On the Reliability of Chinese Economic Data," Journal of Development Studies 12.4 (1976), pp. 438-441; and World Bank, China: Statistical System in Transition (Washington DC: World Bank, 1992).

3. Wu Jiapei, "Economic Forecasting and Policy Analysis," Renmin ribao [People's Daily] 15 June 1982, p. 5, translated in FBIS, 22 June 1982, K6.

4. Of particular note is Ian Johnson's observation that "doubts about the accuracy of. . . [NBS] statistics. . . . have grown over the past three months as the bureau announced a series of feel-good figures for 1998." ("Beijing, Stung by Criticism, Promises to Issue More Accurate Economic Data," Wall Street Journal 25 February 1999, A12).

5. See Thomas G. Rawski and Robert W. Mead, "On the Trail of China's Phantom Farmers," World Development 26.5 (1998): 767-781.

6. Zhongguo tongji nianjian 1981 [China Statistics Yearbook, 1981; Beijing: Zhongguo tongji chubanshe, 1982], p. 129.

7. Zhongguo laodong gongzi tongji ziliao 1949-1985 [Statistical Materials on China's Labor and Wages, 1949-1985; Beijing: Zhongguo tongji chubanshe, 1987], pp. 109, 270.

8. 1990 Zhongguo laodong gongzi tongji nianjian [China Yearbook of Labor and Wage Statistics 1990; Beijing, Zhongguo tongji chubanshe, 1991], p. 49.

9. Zhongguo laodong tongji nianjian 1991 [China Labor Statistics Yearbook 1991; Beijing: Zhongguo laodong chubanshe, 1992], p. 22.

10. Zhongguo laodong tongji nianjian 1994 [China Labor Statistics Yearbook 1994; Beijing: Zhongguo tongji chubanshe, 1995], pp. 68-69.

11. Song Changqing, "Survey and Verification of Unemployment Rates," Zhongguo tongji [China Statistics], no. 3 (1996), pp. 31-33.

12. Quan Xianzuo, "Economic Factors Influencing Social Stability," Zhongguo tongji [China Statistics], no. 11 (1996), pp. 16-18.

13. Hu Angang, "Zhongguo de shiye wenti yu jiuye zhanlue" [China's Unemployment Problem and Employment Strategy]. Report no. 1 in a series Zhongguo guoqing fenxi yanjiu baogao [Research Reports Analyzing China's National Condition]. Internet document dated 9 February 1998. Official rate from Zhongguo tongji nianjian 1999 [China Statistics Yearbook 1999; Beijing: Zhongguo tongji chubanshe, 1999], p. 133.

14. Quan Xianzuo, op. cit.

15. For example, Cheng Liansheng and Liu Xuemin, "How to Deal With China's Sixth Peak of Urban Unemployment," Jingji yanjiu ziliao [Economic Research Materials], no. 1 (2000), pp. 3-9.

16. See Yue Zhihong, "Rising Unemployment Rates in the Transition Period: Causes and Responses," Jingji kexue [Economic Science], no. 2 (1999), pp. 12-17.

17. Wu, Yunhe, "Crackdown on Statistics Abusers," China Daily 7 September 1994, p. 3.

18. Wu Yunhe, "Census Nets Critical Info for Industry," China Daily, 3 April 1997, p. 1.

19. Li Dan, "Statistical Puzzles Confronting Regional Governments," Zhongguo tongji [China Statistics], no. 6 (1999), p. 22.

20. For example, Huang Song and Xu Dehua, "Following Enterprise Reform, Statistical Work at Township and Village Enterprises Needs to be Strengthened," Zhongguo xiangzhen qiye [China Township Enterprise], no 11 (1998), pp. 7-8.

21. Sholnn Freeman, "Detroit Urges All to Take Part in Census," Wall Street Journal, 14 February 2000, A2. See also Nicole M. Christian, "All-Out Fight in Detroit to Keep Census Above a Million," New York Times, 2 May 2000, A18.

22. Wang Zhenhua, "Win the Hardest Battle," Xiangzhen qiye [Rural Town Enterprise], no. 5 (1999), p. 7.

23. Zhao Chengxin and Liu Ping, "Expand the Adjustment of Industrial Structure, Stimulate Healthy, Rapid, and Profitable Development of Township and Village Industry," Xiangzhen qiye [Rural Town Enterprise], no. 10 (1999), p. 8.

24. Yue Ximing, "Exaggeration in the Growth Rate Figures," Gaige [Reform], no. 2 (1999), pp. 74-76.

25. Ye Fengxin, "Wind of Falsification and Embellishment: Origins and Responses," Zhongguo tongji [China Statistics], no. 5 (1998), 31.

26. Xu Binglan, "Statisticians Seek Reliability," China Daily Business Weekly15 February 1999, p. 1.

27.

GDP and energy data from Zhongguo tongji nianjian 1999 [China Statistical Yearbook 1999; Beijing: Zhongguo tongji chubanshe, 1999; cited below as "Yearbook 1999"], pp. 57 and 247.

28. Yearbook 1999, p. 545 and Zhang Ping, "Income Differentials, Interest Rates, and Consumption," Caimao jingji (Finance and Economics), no. 8 (2000), p. 19.

29. Industrial output value and physical commodity output for 1997/98 from Yearbook 1999, pp. 424 and 445-446.

30. For detailed examples, see Meng Lian, "Analysis of Economic Conditions and Policies During the Past Several Years," Gaige [Reform] no. 3 (1999), pp. 73-82.

31. my rough translation from Gan Xinmin and Li Tongyin, "To Control Falsification, We Must Control its Foundations," Zhongguo tongji [China Statistics], November 1998, p. 21.

32. Internet translation of "Corruption Should Be Opposed in Statistics Too!" Keji ribao [Science and Technology Daily], 7 March 1999, courtesy of David Cowhig.

33. Ming Pao [Hong Kong], 10 July 1998, translated in BBC SWB FE/3278 S1/1, 14 July 1998.

34. Tong Ji, "Strengthen Work on Sectoral Statistics," Zhongguo tongji [China Statistics], no. 9 (1999), p. 6.

35. Meng Lian, op. cit., p. 78.

36. "Corruption Should Be Opposed," op. cit.

37. Yearbook 1998, p. 347; Yearbook 1999, p. 340.

38. Agricultural output data from Yearbook 1999, p. 382. For classification of the 1998 floods, see Zhongguo tongji [China Statistics], no. 8 (1999), p. 38.

39. For details, see material on my web site: http://www.pitt.edu\~tgrawski.

40. I am indebted to Albert Keidel and Barry Naughton for detailed comment on this point.

41. The fiscal figures in Table 2 are limited to indirect taxes, a category that encompasses the vast majority of China's fiscal revenue.

42. "Official Confident of Reaching Tax Target," BBC SWB FE/3356 S1/2, 13 October 1998.

43. See ibid. and "Tax Office Must Collect 191bn yuan to Meet 1998 Target," BBC SWB FE/3391 S1/2. 23 November 1998.

44. Ibid.

45. Wang Aiqiong and Lung Junpeng, "Serious Village Indebtedness: A Dangerous Signal," Jingji yanjiu cankao [Economic Research Reference], no. 81 (2000), pp. 45-46.

46. "Finance Minister Supports Active Fiscal Policy at Beijing Work Conference," BBC SWB FE/3411 G5, 16 December 1998.

47. Wang Chuandong, "State to Bolster Demand," China Daily 29 April 1999, p. 1; Bu Ran, "Increased Renting Expected," China Daily Business Weekly, 6 December 1999, p. 6; Yearbook 1999, p. 527.

48. "NBS Deals a Blow to Falsification," dated 28 February 2000, retrieved from the web site of Zhongguo qingnianbao [China Youth Daily], www.cyd.com.cn

49. "NBS Chief Takes the Pulse of Statistical Data," retrieved from web site for Renmin ribao [People's Daily], 29 February 2000. www.peopledaily.com.cn; "NBS Chief Makes His Report," retrieved from web site for Zhongguo qingnianbao, 29 February 2000; "Economist Liu Guoguang Argues Against Inflationary Policies, 8 per cent Target," Renmin ribao, 16 November 1998, translated in BBC SWB FE/3401 S1/1, 4 December 1998.

50. Renmin ribao [People's Daily], 29 February 2000, retrieved from web site www.peopledaily.com.cn.

51. Ding Junfa, "Trends in Domestic Commodity Markets," Jingji guanli [Economic Management], no. 11 (1999), p. 8. The figure for 2000 is from Xu Hongyuan, "Expanding Effective Demand Along With Structural Readjustment Can Extend the Economic Upturn," Hongguan jingji yanjiu [Macroeconomics], no. 7 (2000), p. 20.

52. Zhao Huanxin, "Dynamic Fiscal Policy Needed," China Daily 6 March 2000, p. 1. "Nation Moves Boldly Forward" (text of Premier Zhu's annual report), ibid., p. 5.

53. Jean Dreze and Amartya Sen, Hunger and public action (Oxford : Clarendon Press, 1989), p. 212.

54. Wang Xiaolu, "System Reform and the Sustainability of China's Economic Growth," Jingji yanjiu [Economic Research], no. 7 (2000), p. 11.