Estimating Cross Province and Municipal City Price Level Differences in China: Some Experiments and Results

Abstract

The estimation of Sub-national purchasing power parities (PPPs) for countries where the regions and provinces have different level of development is fundamental for income, consumption, standard of living real term comparisons, as well as for measuring cross-region welfare inequality. This is even truer for large countries like China, where the above aggregates exhibit great variability among provinces. The aim of this paper is to compute the price level differences, measured by the PPPs, for 31 Chinese Provinces and Municipal Cities, based on a sample of 62 goods and services for the year 2014. To our knowledge, this is the first attempt to do it since many years. After a short review of previous studies on China cross-province and municipal cities price level differences measurement and the illustration of methodology and data used, the results of our elaborations are presented and discussed. Taking Beijing as the base area, there is evidence that the PPP max/min ratio is 1.74, confirming the common belief that China cross-province and municipal cities price levels are significantly different.

Measuring the spatial price differences in China with regional price parity methods

“…As indicated by some studies, in large developing countries such as China, the spatial price difference may be more serious than in small developed countries. However, in the existing literature, only a few studies address the RPPs or sub‐national PPPs in China, which mainly including Li, Zhang, and Du (), Brandt and Holz (), Gong and Meng (), Li and Gibson (), Biggeri, Ferrari, and Zhao () and Almås and Johnsen ().…”

“…Among these studies, Li et al () estimated the PPPs for 2002 by taking Shanghai as the norm for the comparisons, and the basket comprised a sample of 127 goods and services. However, as Biggeri et al () criticised, the methodology employed in Li et al () is not clearly explained, and the names of the formulas are not the names generally used in the price indices in the literature. Brandt and Holz () used the provincial‐level price data of 1990 collected for provincial CPI based on a basket of 40–60 items to estimate the 1990 spatial price indices (SPI) by using the Laspeyres formula.…”

“…Very recently, Biggeri et al () employed the GEKS method to compute PPPs based on a sample of 62 goods and services for 2014. They found that there is evidence that the max/min ratio of the PPPs is 1.74, which supports the common belief that the Chinese price levels across provinces and municipal cities are significantly different.…”

“…Moreover, most of the above‐mentioned works including Brandt and Holz (), Gong and Meng (), Li and Gibson () and Almås and Johnsen () have not computed the SPIs as suggested by the ICP methodology. Biggeri et al () also did not estimate the sub‐national PPPs strictly according to the framework of the ICP, although they calculated the transitive indices with the GEKS method; in addition, the sub‐samples for the analysis at the sector level (such as agriculture and industry) were small, as they themselves confirmed.…”

“…This method requires the price data for a bundle of goods and services of a similar quality to be collected for different regions, but such data are usually not available. According to this method, there are several commonly used index formulas including the Laspeyres index (e.g., Gong & Meng, ), GEKS method (e.g., Biggeri et al, ), Törnqvist price index (e.g., Li & Gibson, ), Hedonic regression method together with a weighted CPD approach (e.g., Aten, ; Aten & Figueroa, ; Koo, Phillips, & Sigalla, ; Slesnick, ), and CPD and index method (e.g., Dikhanov et al, ). Unit value approach. This method calculates RPPs based on the Household Expenditure Survey (HES) data, including the information on the amount of expenditure and quantity and the price of the product; the so‐called unit value can be expressed as a ratio of these two components.…”