Jump to ContentJump to Main Navigation
Multinational Firms in ChinaEntry Strategies, Competition, and Firm Performance$

Sea-Jin Chang

Print publication date: 2013

Print ISBN-13: 9780199687077

Published to Oxford Scholarship Online: January 2014

DOI: 10.1093/acprof:oso/9780199687077.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 23 March 2017

(p.176) Appendix 3 Location Decision

(p.176) Appendix 3 Location Decision

Source:
Multinational Firms in China
Publisher:
Oxford University Press

This appendix explains the location choice decisions made by multinational firms between 1998 and 2009. As explained in Appendix 2, while there are a total of 150, 910 foreign subsidiaries that appear in the annual industrial survey database, this analysis also considers the 83, 657 foreign firms that entered between 1998 and 2009, as the most explanatory variables are available during this time period. Our sample firms face a set of location choices, each of which has different attributes. In this model, multinational firms choose which city or region to enter among 22 provinces, four major cities, and five autonomous regions. Thus, each firm has 31 options when choosing a specific location through which to enter China.

The location choice decision can be analyzed with a conditional logit model (McFadden 1974), widely used to understand how individuals or firms choose from a large set of alternatives. Because this model allows us to estimate how each attribute of each region increases or decreases the chance that a location will be chosen over other potential locations, it is applicable to the location choice analysis (Head, Ries, and Swenson 1995; Shaver and Flyer 2000). Since there were 31 regions from which firms could enter during our time study period, each of our 83, 657 sample firms has 31 rows of data, with each row corresponding to a specific region. The maximum likelihood method is used to estimate coefficients to test whether various explanatory variables significantly affect the probability that one region will be chosen among all the regions in the choice set. (p.177)

There are nine regional attribute variables. Population, measured in millions of people, and per capita income, defined as the regional level GDP divided by population in the region, capture the attractiveness of regional markets. The larger and the richer the market, the more attractive it is. The marketiziation index, constructed by the National Economic Research Institute (NERI) (http://www.neri.org.cn) annually, including the years between 1998 and 2009, captures the progress of marketization along different dimensions in each of the 31 provinces in each year. Its major categories include: 1) relation between government and market; 2) development of the non-state sector; 3) development of the product market; 4) development of the factor market; and 5) development of market intermediaries and the legal environment. Three other variables capture institutional barriers: SOE sales share, defined as the proportion of sales of SOEs to total industry shipment in each region, government size, defined as the proportion of state employees to total population in each region, and government subsidy, defined as the proportion of regional government subsidy to total fiscal expenditure in each regional government. We obtain these variables from the Chinese Annual Statistics Yearbook. Horizontal agglomeration, downstream agglomeration and upstream agglomeration are measured by the ratio of firm sales in the same region and in the same 3-digit SIC industry, analogous to Equations (1)–(3) in Appendix 2. Unlike in Appendix 2, however, here we measure foreign firm presence in the 3-digit SIC industry and in the same region. In this Appendix, we also define these variables with both local and foreign firms in model (2) and with foreign firms only in model (3). This allows us to observe whether agglomeration effects differ between foreign and local firms.

Appendix Table 3.1 displays the conditional logit models. In model (1), population and per capita income turn positive and significant, suggesting that foreign firms prefer regions with large populations and high per capita income, both of which reflect market attractiveness. Marketization is also positive and significant, suggesting that foreign firms prefer regions with higher levels of market development, which reflects less government intervention and greater reliance on market (p.178)

Appendix Table 3.1. Conditional logit models of location choice

(1)

(2)

(3)

Population

0.174***

0.099***

0.127***

(0.002)

(0.002)

(0.002)

Per capital income

0.120***

0.119***

0.129***

(0.006)

(0.007)

(0.006)

Marketization

0.525***

0.377***

0.405***

(0.005)

(0.005)

(0.005)

Government subsidy

-0.002***

-0.001

-0.001

(0.001)

(0.001)

(0.001)

Government size

-70.564***

-85.717***

-103.825***

(2.112)

(2.176)

(2.245)

SOE sales share

-3.838***

-2.750***

-3.512***

(0.065)

(0.063)

(0.064)

Horizontal agglomeration

0.434

(0.502)

Downstream agglomeration

0.662***

(0.058)

Upstream agglomeration

5.300***

(0.672)

Horizontal foreign agglomeration

-0.402

(0.680)

Downstream foreign agglomeration

1.209***

(0.101)

Upstream foreign agglomeration

5.946***

(0.902)

Pseudo R-squared

0.260

0.288

0.273

Chi-squared (d.f.)

169473.6 (6)***

185141.3 (9)***

175402.5 (9)***

Observations

2,759, 517

2,726, 621

2,713, 587

Note: Standard errors in parentheses.

(***) p 〈 0.01,

** p 〈 0.05,

* p 〈 0.1.

(p.179) price mechanisms. On the other hand, government subsidy, government size, and SOE sales share are negatively signed and significant, suggesting that foreign firms do not favor a location with large government subsidies, large governments, or larger shares of SOEs.

In model (2), we entered three agglomeration variables, capturing agglomeration of both foreign and local firms. In model (3), we entered the same set of variables, capturing agglomeration of foreign firms only. The agglomeration of firms in the same horizontal industry, whether we consider both foreign and local firms as in model (2) or just foreign firms as in model (3), does not have a significant impact on location choice. This result suggests that competition effects may offset any spillover effects. Although there may be some positive benefits from agglomeration, competing directly with competitors in the same local markets may offset any such benefit. On the other hand, agglomeration in downstream and upstream sectors in a given region shows a strong positive impact on location choice. In other words, multinational firms tend to choose regions with well-established suppliers and buyers who will not compete with them but will simply share their knowledge.