As the role of China in the world economy has risen tremendously in the recent decade, attention to and interest in China’s official statistics is also on the rise. Information on social and economic indicators distributed by China’s National Bureau of Statistics (NBS) currently attracts public attention all around the world. However, Chinese official statistics seem to have a serious image problem, as news reports (especially in the Western media) tend to depict Chinese official data as inaccurate and unreliable. Lack of understanding about the official statistics system of China also contributes to wider audience perception that Chinese statistics are confusing and potentially misleading. The recent developments of unemployment statistics in China show that the authorities basically realize the importance of adequate statistical reporting for economic and social policy. But progress in reforming statistical system is slow, as authorities continue to take a cautious approach in disseminating politically sensitive information.
Why Nobody Trusts Official Statistics
Analysis of government data has always posed a challenge for China watchers, but there is no indicator less credible than the official unemployment rate. Indeed, it may seem rather strange to anyone that the official unemployment rate remained so stable as the Chinese economy went through a series of major events, including restructuring China’s state-owned enterprises, rural-to-urban migration, entry into the World Trade Organization, the global financial crisis of 2007-2008, and the recent expansion of college enrollment. Since the year 2002 unemployment figures in China have remained remarkably flat, within a narrow range between 4 and 4.3 percent. A critical stance toward unemployment data is shared by some Chinese scholars as well.
The main point of criticism is basically the following: the official unemployment rate in China is calculated on the basis of administrative rather than survey data. In estimating national unemployment rates, administrative data are considered to be inferior to population survey-based estimates. Moreover, official figures of “registered unemployed” are subject to national rules and definitions specifically linked to each country’s tradition, so these figures cannot be used for cross-country comparisons. For example, in China’s official statistics the term “registered unemployed persons” refers to the persons in urban areas “with non-agricultural household registration at certain working ages (16 years old to retirement age), who are capable of working, unemployed, and willing to work, and have been registered at the local employment service agencies to apply for a job.”
The low quality of the “registered” urban unemployment rate (RUUR) in China has long been aggravated by inadequate counting of urban migrants. A large fraction of the migrant population lacks local household registration status and thus a lot of those unemployed persons are not qualified to register with local agencies. The situation has started to change recently due to central government steps aimed at relaxing rules on the registration of rural migrants. But so far these policy changes have not had any significant effect on the level of RUUR. Local authorities are slow to implement central government policy decisions, and even those people who are qualified to register may lack knowledge about labor laws or incentive to apply because of low levels of unemployment benefits. According to estimates provided by Chinese scholars, the registration share among the qualified unemployed is very low: in 2001 only 14 percent of unemployed were actually registered and in 2007 this proportion rose to 22 percent.
Alternative Estimates of China’s Unemployment Rate
As the official unemployment rate is widely considered to be uninformative, it comes as no surprise that many of researchers and analysts have attempted to estimate China’s “actual” unemployment rate. Some scholars back in the 1990s came up with an idea to slightly adjust official RUUR figures using information from other official sources. Restructuring of state-owned enterprises at that period led to the phenomenon of laid-off (xiagang) workers, who were not classified as unemployed, but still were counted as a different category by official statistics. Consequently researchers adopted the technique of estimating total unemployment figures by adding together the number of laid-off workers and registered unemployed persons. But the “adjustment” approach has serious shortcomings, as official numbers by definition excluded the large numbers of unemployed without official status and some proportion of xiagang workers were actually employed in the informal sector.
The second approach was developed when researchers managed to get their hands on various arrays of survey data on the Chinese population. These studies usually employed micro-level data collected by NBS or other government related institutions such as the Chinese Academy of Social Sciences. The main drawback of this approach is that survey data are only available for a limited number of regions and for a few number of years. The most recent estimates available date back to the year 2009. Owing to different data sources and methodologies used, existing alternative estimates of unemployment rate also vary greatly, from 6 to 10 percent and even higher.
Given that there is possibly no way to fix the official statistics data, one last option for China watchers is to turn to indices compiled by private sources. For example, Chinese online recruitment platform Zhaopin and the China Institute for Employment Research (CIER) at Renmin University since 2011 have released the CIER Employment Index, which is calculated by dividing the number of job vacancies during a specified period by the number of unique job seekers that apply to jobs during the same period. Therefore, a falling index may signal worsening employment prospects for job seekers. Data are also available for different sectors of the economy, groupings of region and cities, and size of companies. China’s biggest search engine Baidu in mid-2016 announced plans to release indices tracking the state of the economy (and employment in particular) on the basis of information about online requests made by internet users. Of course, online statistics are not free from errors and biases, and private indicators should also be used with caution. But in the uncertain world of Chinese statistics there is certainly no lack of demand for new ways of data measurement.
The Situation Is Improving
Critical assessments of unemployment statistics are justified, but if one takes a broad look at how things have changed over time, signs of improvement are also evident. The very term “unemployment” first appeared in official statistics publications only in the year 1994; before that the Chinese authorities preferred more neutral terms such as “urban population waiting for employment.” As early as 1997 government agencies undertook the first steps to implement survey-based methods for estimating unemployment. The basis for a so called “dual system” was established. Under this system, the Ministry of Human Resources and Social Security was left responsible for providing national unemployment statistics on the basis of administrative data from local employment agencies, and NBS was supposed to provide survey-based unemployment data. However, the first official labor force surveys were conducted only in 2005. Since the year 2010 statistical authorities have expanded geographic coverage of the survey to monitor labor market developments in 31 provincial capital cities, but remained basically silent about results of their pioneering research efforts.
The fog surrounding official unemployment statistics has started to disperse recently, albeit rather slowly. Under the new leadership of President Xi Jinping, the party and central government pay more attention to economic and social stability issues, and consequently monitoring developments in related areas is getting more important than ever. In July 2014 the State Council announced that after several years of research and exploration the authorities were now ready to publish survey-based unemployment estimates to enhance the macroeconomic decision-making process. Since then senior government officials and statistical authorities have begun to cite survey-based estimates of unemployment in their speeches. According to recent official comments, the survey-based unemployment rate stands at around 5 percent.
Despite this fact, NBS still does not publish survey-based estimates on its website in a timely and consistent manner. It was rumored that publication was actually postponed until the end of the 12th Five Year Plan (which covered the period from 2011-2015) in an effort to avoid confusion with previously announced official targets for RUUR. But eventually it seems Chinese authorities have come to the conclusion that the national statistical system does not have capabilities yet to provide adequate survey-based estimates. For the year 2016 a numerical target of keeping RUUR within 4,5 percent was set in the new Five-Year Plan unveiled in March 2016. According to this planning document, additional efforts should be made to enhance the system of unemployment statistics, to establish a sound unemployment early warning mechanism, and to publish survey-based estimates of the urban unemployment rate.
Provincial-level Data Are Needed
The ability of statistical authorities to organize labor force surveys on the regional level is the key to providing reliable estimates of the national unemployment rate in China. Obviously, it is not an easy task for such a huge and economically diverse country as China. But without establishing a country-wide monitoring system, it does not make sense to calculate any national measures of unemployment. The expansion of the regional scope of the NBS labor force survey can also provide more information for authorities and experts about what is going on local labor markets. The rebalancing of the Chinese economy can have a disproportionate influence on different industries and provinces, and currently major fears of a hard landing and unemployment crisis are materializing in the northeastern part of the country.
What China actually needs in these circumstances is not the national average unemployment data but regional unemployment rates. NBS’ decision to expand the coverage of its labor force survey from 31 to 65 cities can be considered a step in that direction. But China still has a long way to go in building a data collection system, ensuring quality and data consistency, and constructing relatively long time-series data on China’s labor market of 800 million workers. So far, the developments in official statistics can be summarized as a mix of good intentions with slow-paced improvements and a lack of transparency.
Dmitriy Plekhanov is a Senior Researcher at the Institute for Complex Strategic Studies (ICSS), a Russian-based think tank focusing on policy-relevant research, providing policy advice to government and business, and facilitating dialogue between business, government, science, and education.