Border Effect Between Hong Kong and Mainland China

• Published date: 01 February 2016
• Author: Jia Wu,Simon C. Fan,Xiangdong Wei
• Date: 01 February 2016
• DOI: http://doi.org/10.1111/1468-0106.12148

BORDER EFFECT BETWEEN HONG KONG AND

MAINLAND CHINA

SIMON C. FAN Lingnan University

XIANGDONG WEI Lingnan University

JIA WU*Jinan University

Abstract. This study applies newly developed methodologies to investigate the border effect between

Hong Kong and mainlandChina. It makes three contributions to a better understanding of the border

effectpuzzle. First, the specialrelationship betweenmainland China and HongKong provides us with a

uniqueopportunity to disentanglethe factors underlyingthe observed border effect.Second, in addition

to estimatingthe border effect betweenmainland China and Hong Kong,we also estimate that between

the mainland provinces of Guangdong and Shanghai. This comparison between two types of border

effects facilitates a better understanding of the extent of domestic market integration inChina. Third,

owing to the nature of our data and the new estimation approach we adopt, our estimated border effects

are less subject to the measurement error-related and estimation-related biases commonly seen in previous

studies. We find the estimated border effect between Hong Kong and mainland Chinato be approximately

266 km, which is substantially larger than that between Guangdong and Shanghai (1.3 km).Both

are substantially smaller than the estimated border effects found in many previous studies.

1. INTRODUCTION

Since the publication of the seminal paper of Engel and Rogers (1996), a large

stream of research testing the effects on trade of national borders has appeared.

Border effects estimated following Engeland Rogers’method are often extremely

large (see Parsley and Wei, 2001),1leading to debate among economists about

whether the existence of political borders and trade barriers can result in such

effects (see Obstfeld and Rogoff, 2000; Ceglowski, 2006).2

The existing literature highlights at least three issues that may compound the

estimated border effect. First, this effect may have numerous causal factors, ranging

from tariffs, exchange rate volatility, and differences in regulatory policies,

transportation infrastructure and wages to consumer tastes (see Anderson

and van Wincoop, 2004; Gopinath et al., 2011). It is therefore important to

identify the influence of different factors, particularly trade versus non-trade

costs. One way to isolate some of these factors (such as currency barriers and

*Address for Correspondence: Department of Economics, Jinan University, 601 Huangpu Avenue

West, Guangzhou, China, 510632. E-mail: jwu@jnu.edu.cn. We are grateful to the Hong Kong Re-

searchGrants Council (ProjectCode: LU341207)and the Guangdong CollaborativeInnovationCenter

of IndustrialTransformationand Upgrading for their financialsupport. We would alsolike to thank an

anonymous referee for the usefulcomments and suggestions thathelped to improve the paper.

1For example, Engel and Rogers’(1996) estimation indicates that the cost of the US–Canada

border is equivalent to a distance of 75 000 miles. Parsley and Wei (2001) apply these scholars’

method to investigate the prices of 27 goods traded in 96 US and Japanese cities, and find the

US–Japan border to be equivalent to a distance of 43 000 trillion miles.

2Ceglowski (2006) reveal that crossing the US–Canadian border adds as much to relative price vol-

atility as an additional distance of 3000–4000 miles. Obstfeld and Rogoff (2000) list the ‘border effect’

as one of the main puzzles in international macroeconomics and one that deserves further research.

Pacific Economic Review, 21: 1 (2016) pp. 3–12

doi: 10.1111/1468-0106.12148

© 2016 John Wiley & Sons Australia, Ltd

bs_bs_banner