For most of the reform era in China, GDP per registered population was reported. This differs from the resident population because of the non-hukou migrants — people that move from their place of registration. There were fewer than five million non-hukou migrants when reforms began in 1978, but now this figure has reached more than 200 million.

At the time of the 2000 census, Guangdong province had a registered population of 75 million but a total resident population of 86 million — the hukou count overstated GDP per capita by 15 per cent. For much larger areas, such as individual counties and big cities, the error is much larger. The city of Shenzhen provides an outstanding example: while its registered population was just over one million at the time of the 2000 census, its total resident population was seven million, so per capita GDP was overstated by almost 600 per cent in the official data.

GDP per capita is most understated in interior provinces like Henan and Sichuan, where the outflow of non-hukou migrants is highest. Conversely, in the main destinations per capita GDP is overstated by an average of 15 per cent (as of 2005). This includes provinces on the southeast coast, such as Shanghai, Zhejiang, Fujian and Guangdong, and the Beijing–Tianjin conurbation. In terms of growth rates, using GDP per registered population incorrectly overstates 1990–2010 trend growth rates for Beijing and Shanghai by about two percentage points per year.

Another problem is the sharp discontinuity created when provinces switch from the hukou count to the resident count for reporting GDP per capita. There is also a double count of up to 26 million people because some provinces switched to the resident count 18 years before the latest province switched. So during that period, someone could be in the denominator of GDP per capita for two provinces at once, as a resident of one province and in the registered population of another slow-to-switch province.

A recent working paper describes these errors in China’s provincial population data. The paper shows that the apparent rise in interprovincial inequality, and the change in trends in regional inequality around 2003, is a statistical artefact resulting from these errors. The perception of rising regional inequality distorted public debate during the reform era, when the opposite is true. The spurious change in trend coincides with initiatives to reduce regional inequality that saw more than one trillion yuan (US$157 billion) invested into infrastructure development in the western provinces. It is difficult to evaluate this investment because of the measurement error the changing population denominators have caused.

These errors have an impact on the trends in interprovincial inequality. Inequality in reported GDP per capita rose at a two per cent annual rate in the decade from 1990, almost double of what the revised resident count, which relies on the 2010 census data, shows. If GDP per capita calculations did not change, the apparent interprovincial inequality would have continued rising sharply. The official GDP figures switched from using registered population to resident population in 2003.  Since inequality in GDP per resident is much lower than inequality in GDP per registered population, this switch automatically reduced measured inequality in the officially reported GDP per capita.  This gives the appearance of a sharp reversal in 2003, without any underlying change necessarily having occurred in the economy.

The changing pattern of inter-provincial inequality in the reform era has four episodes, only one of which involves rising inequality. During 1978–90 interprovincial inequality declined almost continuously. About one-third of this decline was reversed over the next three years and then a one-year rise in 2005 ended a decade of little change in inequality. But even with that rise, inequality returned to only two-thirds of its starting value in 1978. Interprovincial inequality fell sharply after 2005, so that by 2010 it was right back at the low levels previously seen in 1990. The only sustained episode of rising interprovincial inequality was from 1990 to 1993, representing just three years out of the three decades of China’s reform era.

John Gibson is Professor of Economics at the Waikato Management School, the University of Waikato.  He is the co-author, with Chao Li, of the working paper Rising Regional Inequality in China: Fact or Artefact, available here.