China has ordered its National Bureau of Statistics to fight fake data with modern technology, under a revised mandate endorsed by Beijing. The South China Morning Post reported on 8 November that the bureau has been tasked with uncovering fabricated economic data, in the latest update to the agency’s responsibilities. The State Commission Office for Public Sector Reform published the revised scheme on Friday, setting out the bureau’s functions, internal structure and staffing quotas, and the scheme took effect on 28 October.
The office that published the scheme is a joint Communist Party and state body that finalises the roles of government agencies, so its decision carries weight inside the system. The revised mandate is meant to strengthen the bureau’s oversight of data quality and its hunt for fake data with technological tools, though no specific technologies were named in the published notice. The report did not say whether the tools would include artificial intelligence, big data analysis or automated cross-checks against other records.
The bureau sits at the centre of China’s economic reporting. It publishes the country’s closely watched gross domestic product figures, along with data on industrial output, retail sales, investment and employment, and those numbers shape decisions by investors, companies and foreign governments. Doubts about their accuracy have followed the bureau for years, and Beijing has repeatedly said it wants the figures to be trusted.
Fake data has been a recurring problem at local level. Some regions have created the appearance of economic growth by falsifying figures, and some local authorities have issued secret instructions to intervene in data collection. The pressure to report strong growth has often come from the way officials are judged: promotion has long depended in part on the economic numbers their region produces, which gives local leaders a reason to make the figures look better than they are.
Fake data and the earlier crackdowns
China has tried to tackle fake data before. A 20-clause inspection regulation disclosed in September 2018 made the National Bureau of Statistics responsible for organising statistical inspections and building a system to prevent and punish fraud. Between January 2017 and April 2018, the bureau uncovered 72 key cases of statistical fraud covering 26 provincial regions, 2,051 enterprises and 2,942 fixed-asset investment projects, according to China Daily.
The province of Liaoning admitted in January 2017 that economic data from its cities and counties had been falsified from 2011 to 2014. A draft revision to the statistics law put out for public comment in 2019 proposed promoting the use of the internet, big data, cloud computing and artificial intelligence in statistical work, and holding officials accountable when fudged figures were found in their departments or regions.
Fake data checks and what comes next
The new mandate does not by itself say how the bureau will use technology to catch fake data, and the published notice gave no timetable for new systems or new staffing. The scheme sets out functions, structure and headcount, which suggests the change is administrative as much as technical: the bureau has been told that finding fabricated figures is now a formal part of its job, with the staff to match.
Whether the order changes the fake data problem on the ground will depend on what happens next. Local officials still face the same incentives to report strong numbers, and a bureau in Beijing can only check what it can see. The 2018 regulation and the 2019 draft law both promised tougher enforcement, yet falsified figures kept turning up. The test for this mandate is whether the technology it promises produces cases that are found, published and punished.
For now, the order is a signal of intent from Beijing: the bureau that produces China’s headline numbers has been told to treat fake data as a problem to be hunted with modern tools. The scheme has been in force since 28 October, and the next set of provincial figures will be the first chance to see whether anything has changed.