After a decade of crackdowns on statistical fraud, the accuracy of China's economic data has gotten no better and may be getting worse, official findings suggest.
A report presented to China's top legislature on June 20 found an astonishingly high incidence of false economic data provided by local enterprises and officials to the National Bureau of Statistics (NBS).
The findings may raise doubts about the true strength of China's economy as it enters the early rounds of a trade war with the United States.
Bloomberg Economics has estimated that the first tit-for-tat tariffs affecting U.S. $34 billion (225 billion yuan) of trade on both sides would lower China's gross domestic product by 0.2 percent points, making only a slight macroeconomic dent.
"I see minimal macro shock at this stage," Scott Kennedy, deputy director of China studies at the Center for Strategic and International Studies in Washington said by email on Friday.
If retaliatory steps escalate, GDP growth could be cut by half a percentage point, Bloomberg News said last week.
But during China's unacknowledged slump in 2015, economists estimated that the NBS was overstating GDP by a far larger amount, as much as 2 to 3 percentage points, as it reported steady growth of 6.9 percent.
After recovery, China posted the same rate for 2017. The official rate so far this year remains nearly unchanged at 6.8 percent, posing questions about the data behind such rock-solid results.
The NBS's own investigations inspire little confidence in the numbers at a time when China's economic strength may be tested.
Since 2017, the NBS has inspected 2,051 companies and 2,942 fixed-asset investment projects with "abnormal data," state media reported.
The agency found that figures were "severely fabricated" at 1,195 of the companies, or 58 percent of those inspected, the official English-language China Daily said.
The investigation exposed even higher falsification rates among the investment projects with serious fraud in 2,775 cases, or 94 percent, of those examined.
The findings by the NBS Statistical Law Enforcement Supervision Bureau, presented to the Standing Committee of the National People's Congress (NPC), were also notable for the size and degree of the violations.
Companies in the privileged Port Economic Zone of northeast Tianjin city's Binhai New Area were found to have inflated their data by an average of 56 times the real levels, the inspectors said.
Inner Mongolia's Kailu County pumped up its data by a factor of 10, while enterprises in Xifeng County in northern Liaoning province multiplied its results by 6.7 times, according to the report.
Stand-out findings
While cases of data-faking in industrial production figures have become common over the years, the latest findings stand out in several respects.
The persistence of falsification points to either local defiance of the central government or inability to fix the system despite countless efforts and threats to implement reforms.
The report on industrialized Liaoning province, for example, comes a year and a half after the governor, Chen Qiufa, admitted that fiscal data was inflated from 2011 through 2014.
The continuing problem in the province may have special significance due to Premier Li Keqiang's frequently-quoted comments on official statistics as "man-made" and "for reference only" in 2007, when he served as Liaoning's Communist Party secretary.
Li may have unwittingly raised the pressure for overstating gross domestic product figures in Liaoning and neighboring Jilin and Heilongjiang provinces during a 2015 visit, when he blasted officials for letting equipment stand idle, not contributing more to the national economy, and raising the risk of social instability.
Similarly, the new report on Tianjin followed one day after Vice Premier Han Zheng conducted an inspection tour of the city, urging the Binhai New Area to "make greater contributions" to development in the region that includes Beijing and Hebei province.
Whether this was a case of bad timing in light of the falsification report or a downplaying of its implications remains to be seen. People's Daily had already reported in January that the special zone's GDP in 2016 was overstated by about one-third.
Also in January, Inner Mongolia cut its industrialized output figure for 2016 by 40 percent due to earlier falsifications, the official Xinhua news agency and Reuters said at the time.
The continuing and perhaps even greater violations in Liaoning, Tianjin and Inner Mongolia raise doubts about whether the central government's many crackdowns are serious, or compromised by its own drive for growth.
Whatever the answer, the new report shows no sign that enforcement has reduced data fraud at the grassroots level.
"Why does the central government believe that companies and government offices that have never submitted accurate information even know how to do so?" said Derek Scissors, an Asia economist and resident scholar at the American Enterprise Institute in Washington.
"They certainly don't know how to measure return on investment, since the point of investment until now has never been to generate a positive return," Scissors said.
Previous NPC reports and audits over the years suggest that local officials and enterprises may be stuck in a trap of past data inflation, making it difficult to report actual production or profits that are much lower all at one go.
The findings may be a sign that local officials have continued the practice of "smoothing" year-to-year results, known as convergence, despite Beijing's attempts to enforce the Statistics Law, which was last revised in 2009.
"Some ordered companies to fill in reports in accordance with their designated data. Some replaced enterprises' reports with false data. And some even fabricated companies and investment projects," said the Standing Committee's vice chairman, Wang Dongming, according to China Daily.
‘Stability, not accuracy’
The persistence of data fraud is a measure of the political influence over economic policy.
"The main goal of Chinese data generation is stability, not accuracy," said Scissors.
"The numbers are distorted more when the economy is unstable and less when it is stable, so there is no reason to expect continual improvement over time. And there is no way for statistics quality to improve until the Party indicates it will accept reports of instability," he said.
The problems also suggest that high economic growth rates continue to be a main factor for official promotions, despite more recent policies supporting "green" or "high-quality" growth.
The report casts a cloud over the most recent GDP claims and rising returns in the face of international headwinds.
On June 24, for example, the Ministry of Finance reported that five-month profits of the centrally-controlled state-owned enterprises (SOEs) soared 20.9 percent from a year earlier to 1.29 trillion yuan (U.S. $198 billion), topping previous gains in the heavily-indebted sector.
But over the years, officials have been careful to avoid linking the frequent fraud findings to doubts about national economic performance.
In December 2014, the NBS announced it would no longer rely on local GDP estimates after totals from China's 31 provinces, municipalities and autonomous regions repeatedly exceeded the national GDP numbers, in some cases by more than 12 percent.
The NBS claimed the "sole right to evaluate GDP and growth rates of every province and municipality." But since then, the agency has never explained how it reconciles local data with national results.
State media reports have also offered no explanation of why all previous efforts to improve accuracy and reduce fabrication have failed.
In 2012, the NBS announced a direct reporting system for 700,000 enterprises to bypass local official pressure for data inflation. Instead, it appears that local officials simply short-circuited the system by intervening directly with companies to fake their reports.
The new report is strikingly similar to an NPC investigation in 2009 that found rampant fabrication with some production figures overstated by a factor of 10.
But nearly a decade later, the report gives no assurance that the "unified system" reforms for GDP calculations announced in 2014 would be implemented anytime soon.
Last October, an NBS official told Xinhua that the task of taking over data collection from local authorities would be completed in 2019, although it was originally expected in 2016.
State media reports on the NPC findings have not mentioned the unified system or the revised deadline, raising questions about further delays.
‘Checking the checkers’
One potential complication in interpreting the NPC report is that the inspections were conducted by the new Statistical Law Enforcement Supervision Bureau created last year under the NBS, along with similar agencies at the provincial level.
The new supervisory bureaucracy, strengthened with a constitutional amendment at the insistence of President Xi Jinping for all levels of government, supplements the existing anti-corruption machinery, essentially "checking the checkers."
The NBS bureau is a response to Xi's call in May 2017 for a "national supervisory system that will oversee all state organs and civil servants."
The added inspection layer could be catching more cases of data fraud, but supervisory officials could also be magnifying the prevalence of violations to demonstrate their utility and power.