The National Bureau of Statistics (NBS) publicized its methods for compiling the Gini coefficient over the weekend and admitted that it is hard to get access to some types of household income, following queries about the accuracy of its data by some academic institutions and media reports last month.

By December last year, the NBS had collected samples of 140,000 urban and rural households from 31 provinces, autonomous regions and municipalities involving 896 counties and 476 cities and towns, according to an article published Friday by Wang Pingping, director of the households' investigation office under the bureau.

To improve the accuracy of the Gini coefficient, which measures income inequality, the NBS set up a new sample system and began to collect samples from 400,000 households starting from December 1, 2012, the article said.

The specific calculation formula, sample collection procedures and measures were also laid out in the article.

"The calculation formula is okay, but some important details of the original figures for the calculation have still not been disclosed," Gan Li, the director of the Survey and Research Center for China Household Finance under the Chengdu-based Southwestern University of Finance and Economics, told the Global Times Sunday.

"The bureau needs to disclose the percentage of interviewees who declined to receive the NBS' investigations during the sample collection procedure," Gan said. Gan's institute produced its own Gini coefficient in 2010 with a reading of 0.61.

A figure of zero represents equally distributed wealth, with a figure of one representing all wealth concentrated in one individual's hands. The recognized global warning level for unequal distribution of wealth is 0.40.

NBS chief Ma Jiantang announced on January 18 that the country's Gini coefficient stood at 0.474 last year, down from 0.477 in 2011 and a peak of 0.491 in 2008.

The NBS figures raised disputes from some academic institutions that had come up with higher Gini coefficients. And some media reports and Internet users called on the government to publicize its calculation methods.

The Institute of Social Science Surveys under Peking University published a Gini coefficient of 0.514 for 2009, and the National Survey Research Center at Renmin University of China calculated an index of 0.555 for the same year.

The NBS confessed that it is difficult to collect certain types of household income data.

"The major issues in Gini coefficient figures are not their calculation formulas, but the problems in data collection caused by increasing forms of gray income," said Zheng Xinye, a professor at the School of Economics at Renmin University of China.

"A lack of data from high-income households also artificially lowers the Gini coefficient, because wealthier people usually do not like to respond to surveys," Zheng noted.

Zheng said the main cause for lower results in the NBS figures is that they used different household samples. "The income of the wealthy households surveyed by the bureau is lower than in those surveyed by the institutions."