Monday, September 6, 2021

"The Soil That Breeds Statistical Fraud Still Exists"

Chinese officials are acknowledging that the country's statistics are riddled with fraud now that Xi Jinping has targeted statisticians in his expanding campaign to purify the Chinese communist system. Chinese statistics in the "new era" will be improved by endless audits, punishment of perpetrators, and replacement of humans with "smart" computers and gadgets when possible.

The attack on statistical fraud kicked into high gear with an August 30 speech by Xi Jinping summarizing his ideas for improving the quality of data in his "Opinions on more effective use of statistical supervision functions." That was followed by a series of articles posted on the National Bureau of Statistics web site explaining a third round of statistical inspections to root out fraud, falsification, and concealment of data. 

The Bureau's director Ning Jizhe issued slogans for statisticians to contemplate: "Statistical data quality is the lifeline of statistical work." and "Truthful correct, complete and timely supply statistical data is the lifelong task of statisticians." The director explained that, "The soil that breeds statistical fraud still exists, and the struggle against unhealthy trends [continues]."

September 2 article from the all-powerful Central Commission for Discipline and Inspection (CCDI) explained that inspections in 19 provinces over the last two years uncovered problems and improved statistical quality. CCDI warns statisticians that a new round of inspections will address local offices that drag their heels in complying with new statistical requirements, failure to punish or prevent statistical violations, and persistent inaccuracy in data. The new inspections will cover statisticians in all localities and departments. 

CCDI described a web of collaborative deceit perpetrated by local statisticians, companies that report data to them, and local officials that pressure them to report flattering data. One example given was a "shocking" arrangement to inflate industrial statistics in one locality of eastern China. Local statisticians ordered companies to inflate their business income if needed so they could keep their status as an "above scale" company in a database that is tracked to monitor industry performance. Companies had to submit preliminary reports to statisticians so they could determine whether the numbers were big enough. Companies failing to meet the threshold for inclusion in the database were threatened with loss of government projects and other benefits. Companies were ordered to erase electronic chats with the statisticians to cover up the collusion. 

CCDI said some local governments pressure companies to report inflated numbers "out of investment considerations," and some statisticians give companies tables with desired targets and ask them to report data that conforms to these targets. Growth rates are specified for certain months to maintain smooth growth in data. CCDI alleges that some local Economic and Information Departments held meetings to arrange the massaging of data for the year and issue "fraud subsidies" to companies. Some enterprises made false reports to meet the "regulatory" standards. Data reported to statisticians was often inconsistent with tax bureau data, CCDI said, and some did not even submit tax records.

Last week's report on the National Land Survey by the Ministry of Land Resources and Statistics Bureau demonstrated fealty to the crackdown by leading off with the the slogan of "truthfulness and accuracy as the lifeline", and a pledge to "[never] waver in the authenticity of surveys...and severely punish false reports." The survey employed "special inspectors" to scrutinize 394 local survey teams, and the process entailed hundreds of meetings and 7 rounds of "check-feedback-rectification-check again."

Frustrated by immoral and incompetent human statisticians, Chinese officials are putting their trust in machinery to collect statistics. The land survey reported extensive use of technology to minimize "the degree of human interference" in the statistical process. Data was collected with an “Internet + survey” mechanism involving remote sensing, satellite positioning, geographic information systems, mobile internet, cloud computing, and drones to reduce opportunities for statisticians to fake or screw up the data. 

China's National Animal Husbandry Station (NAHS), an arm of the agriculture ministry, held a March meeting to implement a new auditing requirement for the data its network of town, prefecture, provincial and national veterinarians and technicians who count animals and meat. In July, NAHS held a training meeting where officials from national and provincial offices and six livestock and artificial intelligence companies signed a "cooperative framework agreement for joint construction of a comprehensive livestock industry information platform." Officials in the livestock information systems bureaucracy complained about "islands" in data collection, inconsistent data entry at the grass roots level, and failure of local clerks to use APPs. The new system promises a new "shared" platform with "informatized" and "smart" livestock data monitoring to open a new chapter in the modernization of animal husbandry and comprehensive revitalization of the countryside in "the new era."

CCDI blames local statisticians, officials, and companies for statistical fraud, implying that central government and party leaders are victims of misinformation. CCDI blames local officials and companies for paying too much attention to inflating indicators to achieve targets, to burnish the reputation of their city or company, or to get loans or investments. The CCDI finds no blame in the system that incentivizes the fraud by evaluating officials based on statistical achievements and success in meeting targets, distributing transfer payments and subsidies to local governments based on statistical indicators, and handing out bank loans through a state monopoly on banking based on statistical reports. 

Industrial data is the most prominent example of statistical data-bloat, but the incentives extend to every area of the economy where statistical targets are handed down rung by rung in the bureaucratic ladder, including land and livestock data. 

  • Under Xi Jinping's rule hysteria about "food security" has prioritized targets to preserve farmland and "create" new land to "balance" land lost to development; indicators of agricultural production--including land statistics--have been included in evaluation criteria for provincial and local officials. 
  • Progress on a poverty-alleviation program handing out huge amounts of money is monitored with statistical surveys.
  • Big money is invested in "constructing high quality cropland", fruit orchards, and rehabilitation of grasslands. 
  • Counties compete to get transfer payments for being major grain, oilseed and pork-supplying counties on the basis of statistical indicators. 
  • In 2019, officials at all levels were given targets for swine inventories and meat output to accelerate the recovery of pork supplies following the African swine fever epidemic that decimated the pig population. Hence, officials have been competing to see who can report the biggest swine numbers since last year. 

It's unclear whether China's new, improved statistics will be released to the public before they have been massaged and manipulated. Xi Jiping's other priorities include doing a good job on shaping public opinion and telling China's story (not necessarily non-fiction).


Godfree Roberts said...

China's stats are as good as Canada's and Australia's. Read

The quality of China's GDP statistics☆ by Carsten A. HOLZ ⁎
Stanford Center for International Development, Stanford University.

Quality of China's Official Statistics: A Brief Review of Academic Perspectives
Dmitriy Plekhanov

Preeti Sharma said...

This Blog is very informative. Thanks for sharing with us
Basmati Rice exporters

Anonymous said...

That paper by the Stanford guy was published in 2014, yet in 2017 the province of Liaoning openly acknowledged that its GDP data was fake during 2011-2014.

So much for the credibility of such "research".