BACKGROUND PAPER OF CHINA FOR OECD SHORT-TERM ECONOMIC STATISTICS EXPERT GROUP MEETING 10-11 SEPTEMBER 2009, PARIS Hao Shenglong National Bureau of Statistics of China This paper gives a brief introduction about the related fields of short-term economic statistics in China and shares our experiences of dealing with the emerged financial crisis. This paper also is expected to help OECD experts to clearly understand the present situation and mechanisms of the government s economic statistics system. Furthermore, it is expected to contribute to enhance the communication and cooperation between the national statistic offices of OECD member countries and China. Even with the economic boom of the past three decades, China still is a developing country. Comparing with OECD countries, there are many weaknesses, and delays in current government economic statistics in China. Therefore, beneficial comments and suggestions to improve statistics in China are appreciated. 1. Short-Term Economic Statistics in China and the current crisis Usually, short-term statistics refers to the investigation frequency that is shorter than a year in national bureau of statistics (NBS) of China, and it includes semiyearly, quarterly, bimonthly, monthly, ten-days, weekly and daily statistical report systems. The main parts of short-term statistics are monthly and quarterly reports. Many of them are released regularly and are available in the website of NBS: http://www.stats.gov.cn/english/. Content is as follows: Quarterly Data Gross Domestic Product (GDP) Labor Remuneration of Persons Employed in Urban Units Number of Persons Employed in Urban Units by Region Labor Remuneration of Persons Employed in Urban Units by Region Output Value of Farming, Forestry, Animal Husbandry, and Fishery Output Value of Farming, Forestry, Animal Husbandry, and Fishery by Region Producer's Price Index of Major Agricultural Products Per Capita Cash Income of Rural Households by Region Per Capita Cash Expenditure of Rural Households by Region - 1 -
Income of Urban Households by Region Monthly Data Growth Rate of Value-Added of Industry Growth Rate of Value-Added of Industry by Region Growth Rate of Value-Added of Industry by Sector Output and Growth Rate of Major Industrial Products Total Volume of Transportation Post and Telecommunications Services Investment in Fixed Assets Investment in Fixed Assets by Industry Investment in Fixed Assets by Region Investment Completed and Growth of Real Estate Development Enterprises Floor Spaces Sold and Sales Price Growth of Commercial Buildings Sales Price Indices of Buildings in 70 Medium-Large Sized Cities Sales Price Indices of Buildings in 70 Medium-Large Sized Cities (90 sq. meters and below) Total Retail Sales of Consumer Goods Consumer Price Index (CPI) by Category Consumer Price Index (CPI) by Region Retail Price Index by Category Consumer Confidence Index Macro-Economic Climate Index Chart of Macro-Economic Climate Index Monitoring Signals of Macro-Economic Climate Index Facing the financial crisis, NBS has enforced the supervising of the economy performance and adopted new policies on the short term statistics, including: a) Adding price survey on main foods and industrial products each ten-days b) Adding price survey on main agricultural products each half month c) Adding the monthly unemployment survey in the large and middle size cities d) Adding the survey of the utilization of 4 trillion-yuan investment which was announced last year by the central government to prevent the financial crisis e) Try to calculate the chain growth rates based in many fields, especially in GDP f) Conduction of a quick survey of cost of real estate enterprise in May 2009 g) Accelerate the economic performance analysis: A crucial analysis report on economy is made monthly, instead of the quarterly report before 2. New approaches to the measurement of short-term economic statistics of services activities, including follow-up on the Index of Services Production Since the share of tertiary industry in GDP is only 40% in China, the statistics of services activities are relatively weak. There is not an Index of Services Production in China. But you can get the growth rate of tertiary industry from the national account statistics or the growth rates of the turnover of retails, catering or hotels. One important reason of scarcity of ISP is that scales of service entities usually are small and the data is hard to collect. For example, - 2 -
the salary of employee work in small private service shops and restaurants was not covered by the regularly statistics report system. 3. Method of chain-linking and benchmarking in short-term economic statistics Chain-linking and benchmarking are now common in the compilation of short-term economic statistics in most national statistics offices and some international organizations, but they are new for us. Only Hong Kong SAR switched their methodology to these new methods and published chain-linking data since 2007. Based on the understanding of national accounts department in the NBS, both aims of chain-linking method and constant-prince method are to eliminate the price impact in the calculation of GDP growth rate. The chain-linking method is more precise, but the biggest obstacle to adopt the new method in China is basic data scarcity. In details, the current statistics of growth rate in agriculture, industry, construction, transportation, post and communication are based on the constant price methods. So applying the complex advanced chain-linking method in China needs to enhance the training of these news methods at first. The communications and cooperation with OECD Statistics Directorate in these fields also are necessary and urgent. 4. Implement of new international classifications into short-term economic statistics Implement of new international classification usually has a shock on the statistical work, so we are being prudent. In the past three decades, there are three changes of Chinese Standard Industrial Classification (CSIC) in 1984, 1994 and 2002. The national standard CSIC (GB 4754-1984) which was established in 1984 met with ISIC rev.2 (1968) only in agriculture part. The CSIC (GB 4754-1994) which was established in 1994 partially met with ISIC rev.3 (1989) in agriculture, construction and banking parts. In order to shorten the distance with ISIC rev.3, the new CSIC (GB 4754-2002) was revised and published in 2002. This time, it was the same with ISIC rev.3 in section and division levels (in the first 2 digits), but was different in group and group class levels. ISIC Rev.4 had been officially released by United Nations Statistics Division on 11 August 2008. The NACE rev1.1 also moved to rev.2 in 2008 and was adopted in the short-term statistics since 2009. In order to raise the statistics level and catch the advanced countries in the world, China pushed the cooperation with United Nations Statistics Division in the field of coming CSIC revising. An international seminar on applications of International Standard Industrial Classification and Central Product Classification will be opened on 14 th September 2009 in Yunnan Province of China, which will be conducted by United Nations Statistics Division and National Bureau of Statistics of China. In addition, China wants to publish their new CSIC in the late of 2010. The principles of CSIC are made under two key rules. Firstly, it must be based on the real situation and satisfy the needs of economic development in China. Secondly, it should be close to the international standard and satisfy the needs of international comparing. - 3 -
5. Development of short-term business demography statistics Hao Shenglong / National Bureau of Statistics of China We do not have the similar business demography statistics in China, such as entry, exit and survival rates of entrepreneurship indicators programmer. Some related information is dispersed in three government agencies in China: The quarterly and annually Labor Force Survey on salary and employment statistics belong to NBS. The enterprise registration and statistics of small private units belong to State Administration for Industry and Commence (SAIC). The unemployment statistics and social security statistics belong to Ministry of Human Resources and Social Security (MOHRSS). For quality, the statistics of NBS are more reliable and timely than others, but still have some demerits. The coverage of regular quarterly Labor Force Survey in NBS only covers about 45% employed persons in the urban, and excludes the employed persons in private enterprises, individuals and self-employed persons which scale is small and the data is difficult to be collected. The huge migrant workers from rural area also are excluded in the regular Labor Force Survey. The financial crisis caused the decline of demand for labor in the city, and it deeply impacts on the employment of migrant workers. NBS did a fast survey of unemployment of migrant workers in 2008 and has enhanced supervising and regular investigation of migrant workers since 2009. If we enhance the coordination and information share between these three government agencies, especially enhance the statistics of SAIC precisely and timely, we think that short-term business demography statistics can be greatly improved. 6. Revision analysis and the Noise vs News angle The concept of Noise vs News analysis was established in 1984. But this method was never applied to test the efficiency of preliminary estimates in China. The preliminary annual GDP and its growth rate are regularly revised one year later, when the new final raw data are available. Based on the final annual revision, the quarter GDP and its growth rate of last year also are revised. We simply think some news come in, but never test the efficiency of the preliminary data. In addition, the current Chinese quarterly GDP are accumulated data from the year beginning. The corresponding quarterly growth rates are based on the same period of last year. So the individual quarterly GDP and the chain growth rate basted on the previous quarter are not available now. We plan to calculate and publish the individual quarterly GDP and the chain growth rate in 2010. 7. Implementation of the new UN sponsored Revision of the manual on index numbers of industrial production China never published the index of industrial production before. But we have the industry statistics for a long time. It covers the economic activities for mining and quarrying, manufacturing and electricity, water and gas supply. Data of industry are collected monthly, - 4 -
quarterly and annually on value-added and output of industry. The growth rate of value-added is available and published regularly, which is calculated on constant price. Considering the quality and the cost of investigation, only industrial enterprises above designated size with annual revenue from principle business over 5 million yuan are investigated. 8. The use of administrative data in improving short-term statistics Administrative data are widely used in the government statistics in China. For example, there are 14 industries in tertiary industry, but there are only 5 of them that have regular statistical reporting systems in NBS. The national accounts data of remaining 9 industries must be based on administrative data, as this is the only available way. In China, almost half of the government statistics information is not collected by the NBS, but from the administration data of other government agencies. For examples, the monetary data are from People s Bank of China (PBC), the financial data are from Ministry of Finance (MOF), the tax data from State Administration of Taxation, the import and export data are from General Administration of Customs, the utilization of foreign direct investment data are from Ministry of Commerce, etc. Even though administration data are rich and can reduce respondent burden, the quality of administration data remains a potential problem. For example, China adopts the population registration policy, but records of some births and deaths can t be updated timely for special reasons and many people register in a place but work and live in other places. All of these reasons cause the administration data from the population registration to not show the exact number and actual location of population accurately and timely. Another example is that of tax administration data. The efficiency of tax collection has improved much in recent years, but is still not perfect yet. Some tax evasions are still available. If we use the tax administration data to do the short-term statistics estimates, it will cause a bias on the results. - 5 -
References Feng N.L. (2009). News from http://news.163.com/09/0807/03/5g36i91u0001124j.html Ma J.T. (2009). Implement Report of Scientific Development Concept of National Bureau of Statistics (unpublished report). National Bureau of Statistics (2009). Statistical Abstract of China. Beijing China Statistical Press. National Bureau of Statistics (2008). China Statistical Yearbook. Beijing China Statistical Press. National Bureau of Statistics (2008). China Statistical Programmes. (unpublished book). National Bureau of Statistics (2008). Notation of Standard Industrial Classification on Economy. Beijing China Statistical Press. Organization for Economic Co-operation and Development. Compilation Manual for an Index of Service production. From http://www.oecd.org/dataoecd/9/55/37799074.pdf Richard McKenzie, Elena Tosetto & Dennis Fixler. Assessing the Efficiency of Early Release Estimates of Economic Statistics. http://www.oecd.org/dataoecd/20/13/41009155.pdf Zheng X.G. & Liu F. (2009). Application of Administration Data in National Accounting. Journal of China Statistics (331), 6-7. - 6 -