Industrial agglomeration and firm size: Evidence from China

Abstract

This paper, by using annual surveys of manufacturing firms from 1998 to 2005 in China, first documents a positive correlation between industrial agglomeration and firm size, which is previously found in developed economies. Next, by using the instrumental variable estimations, we identify that industrial agglomeration has a positive and statistically significant causal impact on firm size. Finally, we find that firms are more likely to become larger by locating with a number of larger firms than with a larger number of firms.

Introduction

Industrial activities are unevenly distributed across space, e.g., manufacturing belt in the United States (Fritz, 1943), blue banana belt in the European Union (Delamaide, 1994), and Pacific coast industrial belt in Japan (Kitamura and Yada, 1977).¹ The agglomeration of industrial activities has significant impacts on firm behavior and firm performance such as productivity (e.g., Ciccone and Hall, 1996, Henderson, 2003), organization of production processes (e.g., Holmes, 1999, Li and Lu, 2009), and innovation (e.g., Feldman and Audretsch, 1999, Carlino et al., 2007).

In two seminal papers, Kim, 1995, Holmes and Stevens, 2002, by using plant-level data in the United States, find a positive correlation between industrial agglomeration and plant size both across and within manufacturing industries. Subsequent studies further confirm this finding by using datasets from other developed economies, e.g., Ireland (Barrios et al., 2006) and Italy (Lafourcade and Mion, 2007). An interesting research question is that whether the same pattern exists also in developing economies, where economic environments differ a lot from their counterparts in developed economies. And more importantly, does the positive linkage between industrial agglomeration and firm size implies that firms become larger by locating in concentrated industrial areas or reflects the self-selection by larger firms into these areas? And how does industrial agglomeration affect firm size? Answers to these questions have important implications for both academic researches and government policies.² However, very few empirical studies have examined these issues. In this paper, we fill the void by investigating empirically the impact of industrial agglomeration on firm size in China.

China presents a good setting to study this topic. Before 1978, China adopted a central planning system, and nearly all economic activities including location choice and production scale were determined by the central government, which was largely influenced by political considerations. Confronted with the poor economic performance, however, China's central government started to reform its economy by gradually introducing private ownership and market competition in the late 1978. The economic reform not only induced a massive entry of privately-owned enterprises and foreign multinationals, but also restored incentives and decision-making powers in state-owned enterprises. As a result, there have been significant changes in the distribution of industrial activities. For example, Bai et al. (2004) find an upward trend of industrial agglomeration in the latter half of 1985–1997, while Lu and Tao (2009) show that the industrial agglomeration continues during the period of 1998–2005. The fast-changing economic environment in China allows us to examine the interaction between industrial agglomeration and firm behavior.

Our dataset comes from annual surveys of manufacturing firms conducted by the National Bureau of Statistics of China for the period of 1998 to 2005. Using Holmes and Stevens (2002)'s specification, we find that firm size is positively and statistically significantly correlated with industrial agglomeration, which is consistent with the findings in the literature. Meanwhile, in terms of magnitude, the estimated coefficient in China (0.411) is very similar to that in the United States (0.436) as reported by Holmes and Stevens (2002).

To further investigate the relationship between industrial agglomeration and firm size, we follow Henderson (2003)'s estimation framework. Both the pooled OLS estimation and the panel fixed-effect estimation show that industrial agglomeration, measured by a firm's total neighboring employment within the same 4-digit industry and same city, has a positive and statistically significant impact on firm size. To identify whether industrial agglomeration has a causal impact on firm size, we use the instrumental variable estimation à la Li and Lu (2009). The instrumental variable estimation results substantiate our early findings, showing that industrial agglomeration causes firms to become large in production scale. We next include two measures of urbanization economies as in Holmes (1999) in the regression analysis to ensure that our results are not driven by urbanization economies (Jacob, 1969). It is found that our findings are robust to inclusion of these two additional measures, though urbanization economies also cast a positive impact on firm size.

To investigate how industrial agglomeration affects firm size, we decompose our measure of industrial agglomeration into two parts: the number and the average size of a firm's neighboring firms. It is found that both the number and the average size of a firm's neighboring firms have positive and statistically significant impacts on the firm's size, whereas the former has a smaller impact than the latter. These results suggest that a firm is more likely to become larger by locating with a number of larger firms than with a larger number of firms.

We interpret our findings as the evidence of localization economies. ³ To lend further support on localization economies, we conduct a regression of firm productivity on industrial agglomeration and find a positive and statistically significant estimated coefficient, which is consistent with the findings in the literature. However, there are other possible interpretations of our findings, such as product cycle theory (Dumais et al., 2002). To distinguish our interpretation from product cycle theory, we exclude newlly-established firms in the regression analysis. Note that the product cycle theory implies that for old firms, those located in low agglomerated industrial areas should not have much difference in terms of size compared with their counterparts in high agglomerated industrial areas. The regression results without new firms show that industrial agglomeration still has a positive and statistically significant impact on firm size, and the magnitude of this impact is even increased compared with the regression results using the whole sample. These results indicate that the product cycle theory is not applicable in our case.

This paper contributes to the literature by identifying a positive impact of localization economies on an important indicator of firm performance, firm size. Firm size is found to play an important role in underpinning the impacts of institutions on economic growth, such as economic institutions (Laeven and Woodruff, 2007) and financial development (Beck et al., 2008). Meanwhile, firm size is found to have significant impacts on a number of important operation decisions, such as investment in R&D (Cohen and Levin, 1989), financing decisions (Barclay and Smith, 1995a, Barclay and Smith, 1995b), managerial compensation (Jensen and Murphy, 1990), required rate of returns (Banz, 1981, Fama and French, 1992), short-run employment fluctuations (Evans, 1987, Dunne et al., 1989, Audretsch and Mahmood, 1995, Davis et al., 1996), and long-run growth (Fukuyama, 1995). In recognition of the importance of firm size, recently there has been a number of studies investigating what determines firm size. These include the development of financial intermediary (Beck et al., 2006) and the quality of legal institutions (Kumar et al., 2004, Beck et al., 2006, Laeven and Woodruff, 2007). While these papers study the impacts of various economic institutions, our focus here is to investigate whether and how industrial agglomeration affects firm size.

The paper is also related to the recent literature on heterogeneous firms. For example, exporters are found to be larger and more productive than non-exporters (Bernard and Jensen, 1995, Bernard and Jensen, 1999, Bernard and Jensen, 2004, Clerides et al., 1998; Melitz, 2003; Melitz and Ottaviano, 2008), and the big region is most attractive for the more productive firms (Baldwin and Okubo, 2006). This paper is to investigate the firm heterogeneity (in terms of size) across geographic space and identify whether and how industrial agglomeration affects firm size.

The remainder of the paper is structured as follows. Section 2 describes data, and Section 3 presents our empirical findings. The paper concludes with Section 4.