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

• DOI: http://doi.org/10.1111/twec.12899
• Published date: 01 April 2020
• Author: Menggen Chen,D. S. Prasada Rao,Yan Wang
• Date: 01 April 2020

World Econ. 2020;43:1103–1146. wileyonlinelibrary.com/journal/twec

|

1103

© 2019 John Wiley & Sons Ltd

Received: 2 July 2018

|

Revised: 9 August 2019

|

Accepted: 2 November 2019

DOI: 10.1111/twec.12899

ORIGINAL ARTICLE

Measuring the spatial price differences in China

with regional price parity methods

MenggenChen1

|

YanWang2

|

D. S. PrasadaRao3

1School of Statistics, Beijing Normal University, Beijing, China

2Institute of Economic & Social Development, Dongbei University of Finance & Economics, Dalian, China

3School of Economics, University of Queensland, Brisbane, Qld, Australia

Funding information

This study is supported by Project of National Social Science Fund of China (19ATJ002).

KEYWORDS

aggregation, basket products, CPD model, regional price parities, spatial deflation

1

|

INTRODUCTION

The measurement of prices has been an important field in economics because reliable price informa-

tion is required in a vast number of areas that range from micro-topics (e.g., welfare comparisons,

inequality measurements and poverty analyses) to macro-topics (e.g., the calculation of real growth

rates). As Aten and Reinsdorf (2010) pointed out, real comparisons of income, expenditure and wel-

fare generally imply two aspects, namely an adjustment of nominal income for price changes over time

and an adjustment for relative price differences across space. These adjustments exist for a set of

geographic entities and time periods, which means that a measure must be consistent across both

space and time. However, theoretical discussions and statistical practice always pay much more atten-

tion to the adjustment of nominal income for price changes over time but ignore the adjustment for

relative price differences across space either intentionally or unintentionally. The International

Comparison Program (ICP) can be seen as an exception that aims to compare the level of economic

activities and relative prices across countries in the world and produces the purchasing power parities

(PPPs), that is, a spatial price index, as the currency converting factors in the international economic

comparison. However, the measurement of intra-country price differences, specifically, the so-called

sub-national PPPs or regional price parities (RPPs),1 is usually neglected.

Obviously, there are many good reasons to suspect that the price levels and cost of living vary

over space (Deaton, 1988). The need for constructing RPPs has been emphasised in the literature

(e.g., Aten, 2006; Kokoski, 1991), but the international practice in compiling the RPPs has been less

developed, and only a few countries have experimentally produced official indices of spatial prices.

1 In some literature, RPPs are also defined as the regional price index (RPI), such as in Almås and Johnsen (2018).

1104

|

CHEN Et al.

For example, Australia is working towards the regular production of these statistics, while the UK

conducted in 2000 a one-off exercise to produce indicative figures on the variation in prices between

regions and published the estimates of regional price level comparisons in 2004. The United States

undertook various works in this area in the mid-1990s and has estimated experimental price level

differentials in recent years. Italy started a research project in 2005, and since then, it has been con-

ducting an experiment to calculate the sub-national PPPs to compare the consumer price levels at a

regional level (Biggeri & Laureti, 2014). Although the importance of regional price differences cannot

be underestimated in the context of spatial economic analysis, technical problems and data shortages

confine the related statistical practice.

The ICP is a worldwide statistical project designed to compare the levels of economic activity and

relative price levels among countries. The technical advisory group (TAG) of the ICP at the World

Bank discussed and stressed the importance of the computation of sub-national PPPs in its meeting

in February 2010 (World Bank, 2013). However, cross-region price comparisons and the estimation

of PPPs at a regional or any sub-national level such as the city level are a very complicated task. Like

statistical agencies in most countries, the Chinese National Bureau of Statistics (NBS) has not pub-

lished a spatial price index that allows cost-of-living comparisons over space until now. Therefore, the

limited evidence on price dispersion across space makes it difficult to distinguish real inequality from

nominal inequality (Li & Gibson, 2014).

Usually, price differences across regions in large developing countries such as China and India are

more significant, which have a non-negligible influence on the analysis of inter-area living standards,

real income comparisons and poverty measurements (Biggeri & Laureti, 2014). China is the largest

developing country in the world, and this economic development shows significant gaps across re-

gions in fact. The prices of goods and services may vary heavily, which should result in a non-ignor-

able price dispersion in different regions. The available literature about sub-national PPPs in China is

rare and usually suffers from some shortages such as outdated data and biased samples.

This paper intends to estimate the RPPs in urban China with a relatively new sample according to

the scientific framework of PPP methodology in the ICP. The country product dummy (CPD) model,

which is proposed in the cross-country context of the ICP, is employed to estimate the sub-national

spatial price differences of basic headings in the intra-country context, while some popular approaches

including the Gini–Éltetö–Köves–Szulc (GEKS), Geary–Khamis (GK) and weighted country product

dummy (WCPD) methods are used simultaneously in the aggregations to check the robustness of the

results. With a sample of price data from 2015 for 140 product items, which are grouped into 28 basic

headings, the RPPs of 31 provincial regions of China are estimated under a three-step procedure.

Then, the results are extrapolated to 2000 and applied to deflate the regional aggregates and measure

the regional inequality. The remainder of this paper is organised as follows. Section 2 reviews the

theory of RPPs and the related literature; Section 3 introduces the methodology in the estimation of

RPPs. Section 4 introduces the data, and Section 5 addresses the estimation of RPPs, followed by an

extrapolation and application of the results. Finally, Section 6 concludes the paper.

2

|

THEORY AND LITERATURE REVIEW

Regional price parities, that is sub-national PPPs, aim to measure the difference in the price levels of

goods and services across regions for a given time period. Along with the development of interna-

tional economic comparisons, especially the ICP, researchers have for a long time increasingly paid

more attention to the measurement of RPPs (Moulton, 1995).

|

1105

CHEN Et al.

2.1

|

RPPs: as a spatial price index

Theoretically, the price levels and the cost of living vary over space, and a higher price level is ex-

pected in more productive, richer economies (Balassa, 1964; Samuelson, 1964). The same pattern

likely holds within an economy because typically, productivity growth is stronger in the tradable

sectors than in the non-tradable sectors. Usually, the consumer price index (CPI) aims to capture the

change of price over time and can be seen as a temporal price index. Therefore, the CPI may be a

deflator in the adjustment of nominal incomes, expenditures and welfare for price changes over time.

RPPs aim to measure the change of price level over regions and can be seen as a spatial price index

(SPI) used in the deflation of nominal incomes, expenditures and welfare for relative price differences

across spaces (Aten & Reinsdorf, 2010).

Coondoo, Majumder, and Ray (2004), Deaton and Dupriez (2011) and Coondoo, Majumder, and

Chattopdhyay (2011) suggested that the assumption of spatial homogeneity is unlikely to be valid in

the case of large heterogeneous countries with diverse preferences. Accurate regional estimates of

output are desired as an indicator of development levels and as a variable used to explain internal mi-

gration, demand patterns, fertility and other aspects of economic behaviour. RPPs can serve as inputs

in economic comparisons among different areas and can play an important role in advocating and in-

creasing the chances for the overall sustainability of the ICP (Dikhanov, Palanyandy, & Capilit, 2011).

However, most of the international income comparisons treat the entire country as a single entity

and neglect the spatial dimension within a country (Majumder, Ray, & Sinha, 2012). The fact is ig-

nored that in large countries, such as in China, India and Brazil, there is a much greater variation in

prices and consumer preferences among provinces or states than among some smaller countries that

figure in the ICP (Majumder & Ray, 2015). The international statistical agencies have spent consid-

erable resources on calculating the PPPs between nations, while not much attention has been paid to

calculating the RPPs within economies (Asian Development Bank, 2008; Rao, Prasada, & Doran,

2010). Obviously, there is a need to make these comparison deflators consistent across space and over

time. Feenstra, Ma, and Prasada Rao (2009) examined ways of obtaining this space-time consistency

in the international context.

Although the literature on price inflation has generally concentrated on temporal inflation, there

is currently some strong evidence on the spatial variation of prices within a country, and examples

include Hill (2004) on the European Union, Gluschenko (2006) on Russia, De Carli (2008) on Italy,

Majumder et al. (2012), Majumder, Ray, and Sinha (2014a) on India, Majumder, Ray, and Sinha

(2014b) on Vietnam, and Mishra and Ray (2014) on Australia. Aten and Reinsdorf (2010) used US

regional price and expenditure data to further explore the space-time consistency question and esti-

mated a number of traditional multilateral price indices for the years from 2005 to 2008 and a set of 38

geographic areas. The Bureau of Economic Analysis (BEA) first estimated the regional price parities

for 38 urban areas in the United States for 2003 and 2004 (Aten, Figueroa, & Martin, 2012). However,

the literature on sub-national PPPs is still quite recent and limited.

2.2

|

Related studies in China

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

(2005), Brandt and Holz (2006), Gong and Meng (2008), Li and Gibson (2014), Biggeri, Ferrari, and

Zhao (2017) and Almås and Johnsen (2018).