Foreign Direct Investment and Wage Inequality: Evidence from China

World Development Vol. 39, No. 8, pp. 1322–1332, 2011 Ó 2010 Elsevier Ltd. All rights reserved 0305-750X/$ - see front matter www.elsevier.com/locate/worlddev

doi:10.1016/j.worlddev.2010.12.006

Foreign Direct Investment and Wage Inequality: Evidence from China ZHIHONG CHEN, YING GE University of International Business and Economics, Beijing, China

and HUIWEN LAI * The Hong Kong Polytechnic University, Hong Kong, China Summary. — This study provides micro-level evidence on the close link between foreign participation and wage inequality. We investigate the wage premium and the wage spillover effect of foreign-invested enterprises in the Chinese manufacturing sector. The results indicate a significant foreign premium in both wage and non-wage compensation. The presence of foreign and Hong Kong, Macao and Taiwan (HMT) investment results in a significantly negative spillover in terms the wage level in domestic firms, and discourages wage growth in such firms. Overall, the evidence suggests that exposure to foreign investment increases inter-enterprise wage inequality. Ó 2010 Elsevier Ltd. All rights reserved. Key words — wage, inequality, globalization, Asian, China

1. INTRODUCTION

adopting the “open door” policy in the early 1980s, China has become the largest FDI recipient among developing countries, and foreign trade and FDI have been recognized as important engines of economic growth. The other side of the coin is that inequality in China has substantially increased as globalization has progressed. This situation provides a good opportunity to examine the close link between FDI and inequality. We find foreign participation to be a contributing factor to wage inequality. First, our results suggest that there are significant differentials in wage and non-wage compensation among firms with different types of ownership. Foreign-invested enterprises (FIEs) offer the highest wages, whereas private firms and collectively owned enterprises have the lowest wage levels. Second, the presence of foreign and Hong Kong, Macao, and Taiwan (HMT) investment has a significantly negative spillover effect on the wage level in domestic firms. Third, wage growth in multinationals is significantly higher than that in domestic firms. Moreover, the presence of FIEs discourages wage growth in domestic firms, and thus enlarges the wage gap between foreign and domestic enterprises. Collectively, this evidence suggests that access to FDI increases inter-enterprise wage inequality. The remainder of this paper is organized as follows. Section 2 provides some institutional background to the host country wage effect of FDI. Section 3 describes the data and provides an overview of foreign participation in China. Foreign wage differentials and spillovers are investigated in Section 4, and Section 5 concludes the paper.

The host country effect of foreign direct investment (FDI) has been studied extensively in the literature. Many developing countries adopt aggressive FDI promotion policies based on the belief that FDI has a beneficial effect on their economy. Such policies must be justified by both theoretical and empirical studies. Current theories, however, yield ambiguous predictions on the effect of FDI. For example, the procompetition effect of foreign entry may be negative in the short run if multinationals reduce the profit margin and market share of domestic firms, but may be positive in the long run if competition from multinationals forces local firms to increase their efficiency. Technology externalities from FDI may enhance the efficiency of local firms by giving them access to the advanced technology (or intangible firm-specific assets) of multinationals through several channels, such as imitation, labor mobility, and vertical links. The empirical evidence is similarly mixed, and offers no consensus on the impact of FDI on the local economy (see, e.g., Brown, Deardorff, & Stern, 2003; Gorg & Greenaway, 2004; Lipsey, 2004). One important issue that arises in this context is the distribution effect of FDI in developing countries. Goldberg and Pavcnik (2007) have shown that most developing countries have experienced increased inequality contemporaneously with globalization. This study focuses on the link between FDI and inter-enterprise wage inequality. Foreign participation may have a composition effect in that foreign wage differentials directly contribute to inter-enterprise wage inequality. Competition and technology externalities from multinationals may also have indirect wage spillover effects. Foreign wage spillovers from different channels have the potential to increase or decrease the wage level in domestic firms, and thus whether there is a net spillover effect is an empirical question. Our study contributes to this strand of the literature by using Chinese firm-level data to examine foreign wage differentials and spillovers. The case of China has unique value for such a study. As the world’s largest transition economy, China has experienced rapid globalization and achieved rapid economic growth. Since

2. THE HOST COUNTRY WAGE EFFECT OF FDI There is a large body of literature on the wage effect of foreign participation in host countries. Such research can be * We would like to thank Oliver Coomes and five anonymous referees for their insightful comments and suggestions. Lai acknowledges the financial support of Research Grants Council of Hong Kong (PolyU5448/10H). Final revision accepted: November 27, 2010. 1322

FOREIGN DIRECT INVESTMENT AND WAGE INEQUALITY

broadly classified into studies on foreign wage differentials and studies on foreign wage spillovers (Brown et al., 2003; Lipsey, 2004). FIEs may offer higher wages than local firms because they are larger, more capital intensive, and more skill intensive than local firms. FIEs may also pay higher wages than local firms for labor of a given quality, for several reasons. First, due to government restrictions or asymmetric information, foreign firms may operate in a restricted or even segmented labor market, and have to pay higher labor costs than local firms to identify and attract qualified workers. Second, internal fairness policies within multinational enterprises (MNEs) may prevent a large wage disparity from emerging between employees of similar quality in different countries, thereby increasing wages in low-wage regions. Third, FIEs may pay higher wages to reduce labor turnover and thus minimize the leakage of intangible, firm-specific assets. Fourth, FIEs may offer higher wages to compensate for the possible disadvantages of employment in an MNE. For example, workers may prefer local firms, and may have to be compensated to overcome this preference. Workers may also face greater pressure and labor demand volatility in FIEs than in local firms. There is overwhelming empirical evidence to support the claim that foreign firms pay higher wages than domestic firms after controlling for firm characteristics such as size, worker quality, industry, and location. The estimated foreign wage premium is 6–22% in the United States (Feliciano & Lipsey, 2006; Lipsey, 1994), 4–26% in the United Kingdom (Conyon, Girma, Thompson, & Wright, 2002; Driffield & Girma, 2003; Girma, Greenaway, & Wakelin, 2001), about 30% in Mexico and Venezuela (Aitken, Harrison, & Lipsey, 1996), and 10– 50% in Indonesia (Lipsey & Sjoholm, 2002, 2004). Chen, Demurger, and Fournier (2005) use a Chinese household survey to investigate earnings differentials across firms of different ownership types and find that FIEs offer much higher wages than domestic enterprises. Recent studies have controlled for both worker heterogeneity and the selection bias of foreign acquisitions. For example, Heyman, Sjoholm, and Tingvall (2007) use matched employer-employee data to control for employee characteristics, and find that the foreign wage premium is significantly reduced as a result. Almeida (2007) finds that foreign acquisitions of domestic firms have a small effect on human capital and on average wages in acquired firms. Girma and Gorg (2007) find substantial heterogeneity in the post-acquisition wage effect depending on the nationality of the foreign acquirer and the skill group of the workers. The more interesting question is how the presence of FDI affects the wage level and wage growth in domestic firms. Positive wage spillovers help domestic firms to catch up with their foreign counterparts, whereas negative wage spillovers enlarge the wage gap between foreign and domestic firms. Theoretical studies of the effect of FDI on the wage level in domestic firms focus on pecuniary channels and technology externalities. The pecuniary channels of FDI refer to competition between multinationals and domestic firms in both factor markets and product markets. First, competition between FIEs and local firms in the labor market may significantly increase labor demand, and thus oblige local firms to increase their wages to attract a better qualified workforce. The opposite effect may occur if FIEs “poach” the best workers and thus lower both the quality of labor and the wage level in local firms. Second, foreign participation in product markets may result in positive or negative competition effects. On the one hand, foreign participation may force local firms to reduce their margins and become more efficient. On the other hand, competition from multinationals may reduce the market share of local firms, which drives such firms under the minimum effi-

1323

ciency scale or even crowds them out, and thus generates negative spillovers. Technology externalities are also important channels of foreign spillovers. Multinationals, which usually carry superior technology, may generate a positive technology spillover to domestic firms through several channels. First, local firms may adopt new technologies introduced by multinationals through imitation (the demonstration effect). Second, local firms may gain access to the knowledge capital of multinational firms through the labor mobility channel, in that workers previously employed by multinationals may transfer knowledge to local firms by switching employers or starting their own firms. Third, multinationals may transfer technology to firms that are potential suppliers of intermediate goods or are potential buyers of their own products (forward and backward links). These technology externalities may positively affect the efficiency of local firms and improve the domestic wage level. 1 Given that the spillovers from pecuniary channels and technology externalities take place at same time, the net wage spillover effect of FDI is an empirical question. The evidence from previous studies is mixed. 2 For example, Aitken et al. (1996) find a lack of foreign wage spillovers in Mexico and Venezuela, but identify some wage spillovers in the United States. Girma et al. (2001) study FDI in the United Kingdom and find no overall spillover effect on wage levels. Bedi and Cieslik (2002) study the case of Poland and find that workers in industries with a greater foreign presence enjoy higher wages and strong wage growth. Driffield and Girma (2003) find a positive wage spillover effect, but note that it is confined to the region in which FDI takes place. Feliciano and Lipsey (2006) find no evidence of wage spillovers from foreign establishments to domestic establishments in the United States. In the case of China, most studies focus on the effects of FDI spillovers on the productivity of domestic firms, and return ambiguous empirical evidence. For example, Hu and Jefferson (2002) find that FDI generates negative productivity spillovers for domestic firms in the electronics industry, but not in the textile industry. Wei and Liu (2006) find that FDI creates positive intra- and inter-industry productivity spillovers. Hale and Long (2007) find no evidence of systematic positive productivity spillovers from FDI. Liu (2008) shows that FDI lowers the short-term productivity level, but raises the long-term rate of productivity growth among domestic firms within an industry. In terms of foreign wage spillovers, Braunstein and Brenner (2007) find that working in both foreign-invested enterprises and provinces with a higher level of FDI results in higher wages than those paid in other enterprises and regions in both 1995 and 2002. Our study complements the previous literature by investigating how foreign participation affects inter-enterprise wage inequality in China. 3. FOREIGN PARTICIPATION IN CHINA China has experienced rapid globalization during the reform period. Both FDI inflow and trade volume have substantially increased since China adopted the “open door” policy in the early 1980s. China has become one of the largest trading nations in the world and the largest FDI recipient among developing countries. Most previous studies on globalization in China rely on aggregate data, with only a few studies using enterprise-level data to investigate FDI. In this study, we describe foreign participation and the characteristics of FIEs in China based on two firm-level datasets.

1324

WORLD DEVELOPMENT

The first dataset comes from the First National Economic Census conducted in 2004. To the best of our knowledge, the census provides the most comprehensive cross-sectional data available on Chinese enterprises. 3 The basic statistics included in this dataset are summarized in the China Economic Census Yearbook (National Bureau of Statistics in China, 2006). We only access the data for the manufacturing sector, and our study is thus limited in scope to the effect of FDI in the manufacturing sector. We divide firms into six ownership categories according to the registration type reported by each firm: state-owned enterprises (SOEs); collectively owned enterprises (COEs); domestic joint ventures (DJVs); private enterprises; Hong Kong-, Macao, and Taiwan-invested enterprises (HMTIEs); and foreign-invested enterprises (FIEs). 4 In China, the public economy comprises SOEs and COEs, whereas the private sector consists of private firms and foreign (HMT) invested firms. DJVs are shareholding companies and limited liability enterprises that have mixed public and private ownership. Of the total of about 1.4 million manufacturing enterprises, the majority are private firms (66.1%). SOEs account for only 1.6%, COEs for 13.4%, and DJVs for about 10.5%. About 4.4% of firms are HMTIEs and 4.1% are FIEs. The share of FIEs of total manufacturing employment is about 11.6%, and the share of FIEs of total sales is 21.9%. FIEs and HMTIEs together account for about one quarter of manufacturing employment and one-third of manufacturing sales. However, there is potential overestimation bias in the HMT participation rate because a significant proportion of HMT capital is “round-tripping” domestic capital that originates in mainland China and flows back as FDI through Hong Kong to obtain the benefits of favorable policies. Xiao (2005) reviews the causes and implications of round-tripping FDI, and estimates the scale of China’s round-tripping FDI to be around 40%, or at least within the range of 30–50%. Our dataset does not contain any information that can be used to identify the proportion of “round-tripping” domestic capital. This will result in an underestimation of the wage difference and wage spillover effects from HMTIEs on domestic firms.

In Table 1, we compare firm characteristics across different ownership categories. There are significant mean differences across the various categories of ownership. 5 The sample mean wage is 9.1. FIEs have the highest average wage level, followed by HMTIEs and SOEs, with COEs and private firms having the lowest wage levels. We use two measures of firm size: total employment and total sales. The sample mean number of employees is about 104, and the sample mean total sales is 7796. SOEs, FIEs, and HMTIEs have the largest average size, and COEs and private firms the smallest. The average firm age is 6.53 years. SOEs are the oldest firms, followed by COEs, private firms, and FIEs, in that order. The average level of capital intensity, which is defined as fixed assets per employee, is 72.78. SOEs are the most capital-intensive firms, followed by FIEs and HMTIEs, with COEs and private firms being the most labor-intensive enterprises. The average return on assets (ROE) is 0.13. Private firms and COEs have the highest ROE, and SOEs, FIEs, and HMTIEs the lowest. Export intensity, which is defined as the ratio of exports to total sales, is highest in FIEs and HMTIEs. This confirms the finding of previous studies that foreign and HMTIEs are more export-oriented than domestic firms (see, e.g., Blonigen & Ma, 2007; Whalley & Xin, 2006). The skill composition of enterprises is measured by the share of employees with different levels of education. FIEs and SOEs are the most skill-intensive firms, and private firms and COEs the least skill intensive. The second dataset used comprises 10-year panel data for Chinese manufacturing enterprises from 1998 to 2007. This dataset comes from the National Bureau of Statistics Enterprise Dataset. The National Bureau of Statistics of China (NBSC) obtains annual reports from most state enterprises and from large- and medium-sized nonstate enterprises (those with sales of more than five million yuan per year). 6 These annual reports contain the firm’s financial statements and some non-financial information, such as the entry date, district code, industry code, and main products of the enterprise. This dataset is used as the basis for compiling basic statistics on the aggregate manufacturing sector that are summarized in the China Statistical Yearbook (NBSC, 1999–2008) and statistics on two-digit manufacturing industries that are summarized

Table 1. Mean comparison of the characteristics of firms of different ownership types Characteristics Average wage

Total

SOEs

COEs

9.10 11.4 8.30 (7.20) (9.92) (42.3) Total employment 52.4 171.1 42.3 (103.8) (229.3) (78.4) Total sales 7796 22,808 4910 (23,156) (45,450) (16,332) Age 6.53 24.92 13.45 (8.14) (18.10) (10.87) Capital intensity 72.78 250.0 72.51 (255.3) (663.2) (279.1) Return on assets 0.13 0.01 0.11 (0.44) (0.26) (0.43) Export intensity 0.07 0.03 0.03 (0.23) (0.26) (0.16) Share of employees with at least 16 years’ education 0.10 0.20 0.08 (0.19) (0.23) (0.17) Share of employees with at least 12 years’ education 0.41 0.62 0.40 (0.33) (0.30) (0.33) Observations 1,322,341 21,016 176,511

Domestic joint ventures Private firms HMTIEs 9.77 (7.20) 87.4 (150.9) 15,310 (35,560) 6.52 (9.08) 99.09 (310.4) 0.09 (0.34) 0.05 (0.19) 0.16 (0.23) 0.51 (0.33) 139,193

8.56 (5.50) 34.9 (63.3) 4625 (13,961) 4.70 (5.46) 53.07 (166.9) 0.17 (0.47) 0.04 (0.18) 0.08 (0.17) 0.37 (0.32) 873,762

12.1 (8.38) 144.0 (185.1) 22,097 (40,182) 6.47 (4.97) 140.3 (415.9) 0.01 (0.37) 0.43 (0.45) 0.15 (0.22) 0.48 (0.32) 57,752

FIEs 14.7 (10.7) 134.6 (182.8) 27,973 (47,530) 5.23 (4.51) 188.4 (507.7) 0.02 (0.30) 0.39 (0.44) 0.23 (0.28) 0.57 (0.34) 54,107

Note: The standard deviations are reported in parentheses. The results of the mean difference test can be provided upon request. The data source is the Economic Census of 2004.

FOREIGN DIRECT INVESTMENT AND WAGE INEQUALITY 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

Year Foreign Share by Number of Firms

Foreign Share of Employment

Foreign Share of Sales

Figure 1. Foreign Participation in China, 1998–2007. Data source: Enterprise Dataset for 1998–2007.

in the China Industry Economy Statistical Yearbook (NBSC, 1999–2008). Figure 1 shows the time trend of foreign participation in the manufacturing sector during the period 1998–2007. We use three measures of foreign participation: FIEs as a share of the total number of firms, foreign firm employment as a proportion of total employment, and the ratio of foreign firm sales to total sales. The patterns of these measures are consistent, and indicate a significant upward trend in foreign participation. During the sample period, the percentage of firms that were foreign firms increased from 17% to 22%. The share of employment provided by foreign firms increased from 14% to 34%, and the ratio of foreign sales to total sales increased from 27% to 35%. These results are consistent with previous findings on the important role of foreign participation in China’s economy (see, e.g., Whalley & Xin, 2006). 4. FOREIGN WAGE DIFFERENTIALS AND SPILLOVERS (a) Foreign participation and wage and non-wage compensation Previous studies conducted in both developing and developed countries have identified the existence of significant foreign wage differentials. However, few studies investigate the existence of a foreign premium for non-wage compensation. In this section, we examine differentials in both wage and non-wage compensation between multinationals and domestic firms. We first follow Lipsey and Sjoholm (2004) by estimating the following wage equation. X cm Ownershipmi þ b1 Sizei þ b2 ðK i =Li Þ wi ¼ a þ X þ b3 Femalei þ b4 Skilli þ dj Provinceij X þ hk Sectorik þ ei ; ð1Þ where wi is the logarithm of the average wage level in enterprise i. Ownership is measured by five dummies: SOE, COE, private enterprise, HMTIE, and FIE. The benchmark is DJVs. These ownership variables capture the wage differential across ownership types. Sizei is the logarithm of the total sales of enterprise i. Previous studies show that size is positively correlated with wages (e.g., Lipsey & Sjoholm, 2004). K i =Li is the capital–labor ratio, which is defined as fixed assets divided by the number of employees in enterprise i. More capitalintensive firms are expected to have a higher marginal product of labor and thus a higher wage level. Femalei is the propor-

1325

tion of female workers of the total number of employees in enterprise i. Previous studies suggest significant male–female wage differentials, and thus the proportion of female workers is expected to be negatively correlated with the average wage level (e.g., Gunderson, 2006). Skilli is the skill composition of the employees in enterprise i, and is measured by three variables: the proportion of employees with a graduate education (18 years of education and over), the proportion of employees with a college education (16 years of education), and the proportion of employees with a high school education (12 years of education). The benchmark is employees without a high school education. Firms with a higher proportion of skilled labor will have a higher average wage due to the skill premium. Provinceij is a province dummy that is equal to one if enterprise i is located in province j, and zero otherwise, and captures region-specific wage differentials. Sectorik is an industry dummy that is equal to one if enterprise i operates in industry k, and zero otherwise, and captures industry-specific wage differentials. a is a constant and ei is the error term. We use cross-sectional data from the First National Economic Census conducted in 2004 to estimate the wage equation. The ordinary least-squares (OLS) results are reported in column 1 of Table 2. The results indicate that there are significant wage differentials across ownership types when other firm characteristics are controlled. FIEs pay the highest price for labor at a given education level, followed by HMTIEs and SOEs, with the wage levels in COEs and private firms being the lowest. The estimated coefficient of the FIE dummy is 0.267 and is significant at the 1% level, suggesting that foreign wages are about 27% higher than wages in DJVs. The wage premium for SOEs and HMTIEs is about 11%. COEs and private firms pay about 6–8% less for labor than DJVs. These results are consistent with the finding of previous studies of the existence of a significant foreign wage premium. For example, Lipsey and Sjoholm (2004) find that in Indonesia, the foreign wage premium in relation to domestic private firms is a little over one quarter for blue collar employees and one half for white collar employees when skill composition and industry and regional effects are controlled. For other enterprise characteristics, firm size and capital intensity are positively related to wage levels, indicating that larger and more capital-intensive firms offer higher wages. The proportion of skilled labor has a significantly positive effect on wages, suggesting that more skill-intensive firms have a higher average wage level. The proportion of female employees is negatively associated with the average wage level. This implies the existence of significant gender wage differentials, with firms with more female employees tending to offer lower wages. This pattern is consistent with the findings of previous studies on gender wage inequality in China (see, e.g., Shen & Deng, 2008). The 2004 Census reports enterprise expenditure in three categories of non-wage compensation: pensions and health insurance, housing subsidies, and labor and unemployment insurance. Pensions and health insurance are mandatory benefits. Labor and unemployment insurance include various direct subsidies to employees and mandatory unemployment insurance. The census provides information on expenditure on labor and unemployment insurance for all enterprises, but information on the other two categories of benefits is available only for a limited sample. 7 In our sample, only 15% of firms provide a housing subsidy to their employees, 55% provide pensions and health insurance, and 25% provide labor and unemployment insurance. We apply two empirical models to examine foreign differentials in non-wage compensation: the first is a logit model that explains whether a firm decides to provide a certain type of

1326

WORLD DEVELOPMENT Table 2. Wage and non-wage compensation differentials: cross-sectional analysis Wage

SOEs

HMTIEs FIEs Size Capital intensity Female share Graduate College High school Industry dummies Province dummies R2 Observations

Pensions and health insurance

Labor and unemployment insurance

OLS (1)

Logit (2)

Tobit (3)

Logit (4)

Tobit (5)

Logit (6)

Tobit (7)

0.110*** (19.88)

0.740*** (31.91) 0.436*** (16.32)

2.856*** (25.13) 1.527*** (12.16)

0.516*** (24.14) 0.154*** (8.51)

3.409*** (29.72) 0.669*** (6.36)

0.754*** (43.95) 0.037*** (4.13)

3.523*** (38.55) 0.085* (1.69)

1.402*** (67.05) 0.220*** (9.04) 0.401*** (19.36) 0.476*** (80.57) 0.0004*** (21.44) 0.427*** (11.44) 2.138*** (8.80) 2.579*** (65.67) 0.959*** (31.78) Yes Yes 0.25 254,636

5.462*** (55.58) 0.883*** (7.47) 1.663*** (16.25) 1.613*** (57.69) 0.002*** (23.76) 1.632*** (9.19) 8.863*** (7.51) 10.707*** (54.69) 3.765*** (25.87) Yes Yes 0.10 254,636

0.605*** (45.97) 0.126*** (6.69) 0.318*** (17.06) 0.387*** (85.24) 0.0003*** (13.57) 0.063*** (2.62) 1.095*** (4.81) 1.505*** (45.42) 0.344*** (16.77) Yes Yes 0.15 254,636

2.878*** (38.28) 0.290*** (2.93) 0.487*** (5.06) 0.960*** (40.06) 0.004*** (45.10) 0.541*** (3.94) 5.866*** (4.82) 6.634*** (36.92) 1.635*** (13.92) Yes Yes 0.02 254,636

0.512*** (70.55) 0.096*** (8.20) 0.287*** (24.38) 0.249*** (125.95) 0.0001*** (9.31) 0.145*** (13.57) 1.171*** (14.53) 1.433*** (103.75) 0.532*** (59.31) Yes Yes 0.13 12,02,144

2.324*** (56.05) 0.058 (0.89) 0.705*** (10.82) 0.819*** (74.33) 0.001*** (29.20) 0.923*** (15.27) 5.093*** (11.03) 6.602*** (83.95) 2.445*** (48.13) Yes Yes 0.03 12,02,144

COEs

Private firms

Housing subsidy

0.077*** (36.40) 0.063*** (36.45) 0.113*** (36.16) 0.267*** (78.81) 0.024*** (48.82) 0.0003*** (48.82) 0.147*** (61.19) 0.669*** (21.73) 0.420*** (97.32) 0.076*** (35.98) Yes Yes 0.17 11,84,597

Note: The t-values are reported in parentheses; The data source is the Economic Census of 2004. * Represent significance at the 10% level. *** Represents significance at the 1% level.

benefit based on its characteristics. The second model is a Tobit model explaining the benefit per employee based on the firm’s characteristics. The control variables are the same as those in Eqn. (1). The estimation results of the logit model are reported in columns 2, 4, and 6 of Table 2, and the results of the Tobit model are reported in columns 3, 5, and 7. The results of the two models are consistent, and indicate that there are significant differentials in non-wage compensation across the various ownership categories. The logit model estimation shows that SOEs and FIEs are more likely to offer non-wage compensation than DJVs, and that COEs and private firms are least likely to offer such benefits. The Tobit model estimation suggests that SOEs provide their employees with the best benefits, followed by FIEs. COEs and private firms provide the worst benefits. (b) Foreign wage spillovers To control for unobserved individual characteristics, we use the National Statistical Bureau Enterprise Dataset to construct a 10-year panel dataset for the period from 1998 to 2007. This includes most state enterprises and large- and medium-sized non-state enterprises that operated in the manufacturing sector in China over that timeframe. We adopt a fixed effect model to estimate the following wage equation. wi;t ¼ a þ b1 Sizei;t þ b2 Capital intensity i;t X þ cm Ownershipmi;t þ kt þ li þ eit ;

ð2Þ

where wit is the logarithm of the average wage level in enterprise i in year t, Sizei;t is the logarithm of total sales of enterprise i in year t, and Capitalintensity i;t is the capital–labor ratio, which is defined as fixed assets divided by the number of employees in enterprise i in year t. Ownership is measured by SOE, COE, private firm, HMTIE, and FIE dummies. a is a constant and kt is a set of year dummies that control for possible variation in the macroeconomic environment over time; li is included to control for the unobservable individual effect of firm i that could be correlated with the wage level; and ei;t is the error term. 8 The fixed effect estimations with heteroskedasticity-consistent standard errors are reported in column 1 of Table 3. 9 The results suggest that foreign and HMT participation has a significantly positive effect on the enterprise wage level, whereas state and collective ownership have a significantly negative effect on the enterprise wage level. In comparison with the cross-sectional results, the foreign wage premium is significantly reduced from 27% to 2.7% and the HMTIE wage premium is reduced from 11% to 2.4%. These premiums remain statistically significant after controlling for the individual firm effect. The possible reason for the smaller premiums in the panel analysis is that the fixed effect model explains withingroup variation, and there is much less within-group variation in ownership than there is between-group variation. FDI in China consists largely of “greenfield investments” as opposed to foreign mergers and acquisitions (foreign M&A). Few foreign takeovers occurred in the early period of economic reform because the Chinese government adopted a very

FOREIGN DIRECT INVESTMENT AND WAGE INEQUALITY

1327

Table 3. Wage level differentials and spillovers: fixed effect model Wage level

Size Capital intensity SOEs COEs Private firms HMTIEs FIEs

Within-industry spillovers

Within-province spillovers

FE (1)

FE with lagged regressors (2)

Total (3)

Domestic (4)

Total (5)

Domestic (6)

Total (7)

Domestic (8)

0.134*** (168.39) 0.001*** (141.15) 0.047*** (14.04) 0.007*** (3.26) 0.002 (1.19) 0.024*** (5.61) 0.027*** (6.17)

0.051*** (55.74) 0.0001** (5.10) 0.037*** (10.20) 0.017*** (6.85) 0.002 (1.03) 0.026*** (5.77) 0.034*** (7.13)

0.134*** (168.19) 0.001*** (141.14) 0.047*** (13.86) 0.007*** (3.10) 0.002 (1.27) 0.024*** (5.66) 0.028*** (6.44) 0.009 (0.71) 0.114*** (11.62) Yes 0.28 1,986,403 (503,580)

0.136*** (149.66) 0.001*** (119.33) 0.036*** (10.34) 0.002 (0.84) 0.001 (0.01)

0.134*** (168.39) 0.001*** (141.12) 0.047*** (14.02) 0.008*** (3.29) 0.002 (1.19) 0.024*** (5.61) 0.027*** (6.17) 0.085*** (5.00) 0.001 (0.07) Yes 0.28 1,986,403 (503,580)

0.136*** (149.97) 0.001*** (119.31) 0.036*** (10.34) 0.002 (0.82) 0.001 (0.14)

0.133*** (166.97) 0.001*** (140.98) 0.043*** (12.67) 0.006*** (2.77) 0.004** (2.06) 0.026*** (6.01) 0.030*** (6.78) 0.109*** (3.23) 0.513*** (24.89) Yes 0.28 1,986,403 (503,580)

0.136*** (148.95) 0.001*** (119.29) 0.034*** (9.82) 0.001 (0.45) 0.002 (1.07)

Share of HMT employment Share of foreign employment Year dummies Within R2 Observations (Firms)

Within-industry and province Spillovers

Yes 0.28 1,986,403 (503,580)

Yes 0.23 1,458,471 (386,519)

0.094*** (5.92) 0.092*** (7.63) Yes 0.28 1,571,021 (417,706)

0.115*** (5.52) 0.031 (1.26) Yes 0.28 1,571,021 (417,706)

0.591*** (14.69) 0.324*** (13.44) Yes 0.28 1,571,021 (417,706)

Note: Column 1 reports the fixed effect estimations with heteroskedasticity-consistent standard errors; column 2 reports the fixed effect estimations with AR(1); and column 3 reports the fixed effect estimations with lagged regressors. Columns 4–9 report the fixed effect estimations of wage spillovers with heteroskedasticity-consistent standard errors. The t-values are reported in parentheses. The data source is the enterprise dataset for 1998–2007. ** Represent significance at the 5% level. *** Represent significance at the 1% levels.

restrictive approach to granting approval for foreign M&A transactions. Due to China’s accession to the World Trade Organization (WTO) and the gradual deregulation of FDI, foreign M&A transactions involving domestic enterprises have increased in recent years, but are still subject to strict regulation. The Chinese government issued the “Provisions on the Acquisition of Domestic Enterprises by Foreign Investment” in 2006, which permit the use of foreign shares as consideration for the acquisition of Chinese enterprises. However, foreign M&A transactions that would result in a foreign entity gaining actual control of a domestic enterprise in a key industry or in a change in control of a famous Chinese trademark or brand require Ministry of Commerce approval. The Ministry of Commerce also has the right to determine whether the domestic acquisition target has been appropriately valued, and to initiate anti-monopoly investigations into foreign M&A transactions. These regulations effectively discourage foreign takeovers. In the random effect model, the estimated foreign wage premium is about 25% and the HMT wage premium is about 19%, figures that are significantly larger than the fixed effect estimations. However, because the Hausman test rejects the null hypothesis that the fixed effect estimation is not significantly different from the random effect estimation, we report only the results of the fixed effect model. One important issue is potential endogeneity bias if multinationals “cherry pick” the best domestic firms through foreign M&A. If this is the case, then the positive link between foreign ownership and wages may be due solely to domestic firms with high wage levels being taken over by multinationals. To address this issue, we first argue that foreign M&A transactions are strictly regulated and discouraged in China, which restricts

or prevents multinationals from “cherry picking” the best domestic firms. Second, we alleviate this potential bias by use lagged regressors to estimate the wage equation. The results, reported in column 2 of Table 3, are consistent with those reported in column 1. In summary, our analysis suggests the existence of a significant foreign (HMT) wage premium. The more interesting question is how the presence of FDI affects wage levels and wage growth in domestic firms. To investigate foreign wage spillovers, we extend Eqn. (2) by including the FDI presence within each industry and region, as follows. wit ¼ a þ b1 Sizeit þ b2 Capital intensity it X þ cm Ownershipmit þ b3 Fsharejkt þ b4 HMTsharejkt þ kt þ li þ eit ;

ð3Þ

where Fsharejkt and HMTsharejkt are the share of employment of FIEs and HMTIEs of total employment in province j and industry k at time t, respectively. 10 The definitions and measurement of the other variables are the same as in Eqn. (2). The results of the fixed effect model are presented in columns 3–8 of Table 3. We investigate wage spillovers in three ways. First, we limit the spillover effect to within the same two-digit industry and the same province. Column 3 of Table 3 shows that the coefficient of the HMTIE employment share is negative but insignificant, and that the coefficient of the foreign employment share is significantly negative. This suggests that increasing foreign employment is negatively related to the average wage level within an industry and within a region. To investigate the effect of FDI on the wage level in domestic

1328

WORLD DEVELOPMENT

firms, we use only the sample of domestic firms to estimate Eqn. (3) and report the results in column 4 of Table 3. The presence of both foreign and HMT investment in the same industry and province has a significantly negative effect on the wage level in domestic enterprises in that industry and province. A 10-percent increase in the foreign or HMT employment share in the same province and industry lowers the wage level in domestic firms by about 0.9%. Second, we assume that foreign wage spillovers operate across regions but are restricted to the same industry. Foreign presence is measured by the share of employment of FIEs and HMTIEs of total employment in the same two-digit industry. Columns 5 and 6 of Table 3 report the results for all firms and domestic firms, respectively. The results indicate that although there are significantly negative wage spillovers from HMT investment, the spillovers from foreign investment are insignificant. The difference in wage spillovers between HMT and foreign investment may be due to different technology externalities. Technology externalities from FDI are strongly dependent on the source country. Previous studies have shown that HMT investment delivers fewer technology spillovers than foreign investment. For example, Wei and Liu (2006) find that OECD-invested firms play a much greater role in technology spillovers than overseas Chinese firms from Hong Kong, Macao, and Taiwan. In our study, the estimated spillover effect is the net effect of pecuniary channels and technology externalities. Positive wage spillovers from foreign investment through technology externalities may offset negative spillovers through pecuniary channels. For HMT investment, the negative competition effect is likely to result in small technology spillovers only. Third, we investigate the within-region wage spillover. FDI presence is measured by the share of employment of FIEs and HMTIEs to total employment in the same province. The estimation results are reported in columns 7 and 8 of Table 3. The results suggest that an increase in the presence of both foreign

and HMT investment significantly lowers the wage level in domestic firms in the same province. The spillover effects are both statistically and economically significant. A 10% increase in the HMT employment share in the same province lowers the wage level in domestic firms by about 6%, whereas a 10% increase in foreign participation reduces the average wage in domestic firms by about 3%. One concern is that foreign spillovers may be confined to a more disaggregated industry level or to a smaller geographic unit. As a robustness test, we use alternative measures of FDI presence to examine foreign wage spillovers to domestic firms. Columns 1 and 2 of Table 4 report the results when we use the share of foreign and HMT employment to total employment within the same three-digit and four-digit industry, respectively. The results show that foreign wage spillovers are insignificant at the three-digit industry level but significantly negative at the four-digit industry level. Thus, the competition effects of multinationals outweigh technology spillovers at a disaggregated industry level. Column 3 reports the results for within-city or within-prefecture foreign spillovers. The results show that the presence of foreign and HMT investment has a significantly negative effect on the wage level in domestic firms. Another concern is potential endogeneity bias. Labor costs are an important determinant of location choice among multinationals. If foreign investors are more likely to enter a lowwage region or industry, then the negative foreign wage spillovers will be overestimated. To address this issue, we use government FDI policy indexes as instruments to measure FDI participation. Previous studies have shown that government preferential policies are an important determinant of FDI location in China (e.g., Amiti and Javorcki, 2008; Cheng & Kwan, 2000; Head & Ries, 1996). For regional FDI participation, we measure government preferential policies by the number of incentive zones in each province, including Special Economic Zones, economic and technological development

Table 4. Wage spillovers to domestic firms: robustness test Fixed effect model

Size Capital intensity SOEs COEs Private firms Share of HMT employment Share of foreign employment Year dummies Within R2 Observations (Firms)

Two-stage fixed effect model

Within three-digit industry (1)

Within four-digit industry (2)

Within city or prefecture (3)

Within four-digit industry (4)

Within province (5)

0.136*** (201.7) 0.001*** (198.0) 0.036*** (11.68) 0.002 (0.91) 0.001 (0.17) 0.080*** (6.22) 0.003 (0.24) Yes 0.28 1,571,021 (417,706)

0.136*** (149.91) 0.001*** (119.33) 0.036*** (10.35) 0.002 (0.83) 0.001 (0.17) 0.044*** (3.84) 0.041*** (3.22) Yes 0.28 1,571,021 (417,706)

0.135*** (148.84) 0.001*** (119.40) 0.034*** (9.79) 0.001 (0.63) 0.002 (1.04) 0.390*** (11.46) 0.335*** (15.41) Yes 0.28 1,571,021 (417,706)

0.126*** (151.87) 0.001*** (153.66) 0.028*** (8.33) 0.003 (1.21) 0.002 (1.12) 0.145*** (4.38) 0.055* (1.67) Yes 0.27 1,147,643 (318,132)

0.126*** (149.30) 0.001*** (152.09) 0.027*** (7.86) 0.005** (2.09) 0.006*** (2.94) 1.368*** (24.80) 0.276*** (9.52) Yes 0.27 1,130,246 (314,015)

Note: The t-values are reported in parentheses. The data source is the enterprise dataset for 1998–2007. * Represent significance at the 10% level. ** Represent significance at the 5% level. *** Represent significance at the 1% levels.

FOREIGN DIRECT INVESTMENT AND WAGE INEQUALITY

zones, and national foreign trade zones. 11 Local governments in China have developed various types of incentive zones and implemented favorable policy for foreign affiliates located in these zones to promote FDI. There are large variations across provinces, with coastal areas having more incentive zones than inland areas. For industry FDI participation, we use industrial FDI policy indexes as instruments. In the mid-1990s, the Chinese government implemented the “Provisional Guidelines for Foreign Investment Projects” to promote FDI according to domestic industrial objectives. The guidelines were adjusted over time to be consistent with industrial policy. FDI projects in different industries are classified into four categories: encouraged, permitted, restricted, and prohibited. We use three dummies—a permitted dummy, a restricted dummy, and a prohibited dummy—as instruments. The benchmark is an industry in which foreign entry is encouraged. We also include the lagged values of foreign and HMT presence as instruments, and assume that these lagged variables are asymptotically uncorrelated with the error term. The two-stage fixed effect estimations are reported in columns 4 and 5 of Table 4. The results show that the estimated coefficients of HMT presence are negative and statistically significant, indicating that the negative competition effect dominates the positive effect of technology externalities. For foreign investment, the within-region spillovers are significant, but the within-industry spillovers are only weakly significant. This implies that technology externalities from FDI are more significant within an industry than they are across industries, and that within-industry technology spillovers are likely to offset the negative spillovers that occur through pecuniary channels.

1329

(c) Foreign participation and wage growth We next analyze how FDI participation affects wage growth. Foreign spillover effects on the wage level and wage growth in domestic firms may be different. In the short run, competition from multinationals may reduce the market share of local firms, attract good workers, and thus lower the wage level in domestic firms. In the long run, the presence of multinationals may improve the efficiency of local firms through competitive pressure and technology externalities. For example, Liu (2008) finds that an increase in FDI presence lowers the short-term productivity level but raises the long-term rate of productivity growth in domestic firms in China. To identify the effect of FDI on wage growth, we use the panel data to estimate the following equation. gi;t ¼ a þ b1 wi;t1 þ b2 Sizei;t1 þ b3 Capital intensity i;t1 X þ cm Ownershipmi;t1 þ kt þ li þ eit ;

ð4Þ

where git is the wage growth rate of enterprise i in year t and wit1 is the logarithm of the average wage level in enterprise i in year t  1. The definitions of the other variables are the same as in Eqn. (2). The results of the fixed effect model are reported in column 1 of Table 5. The estimated coefficient of lagged wage is significant and negative, indicating that wages grow faster in firms with lower wage levels. The results also suggest that wages grow faster in larger and more capital-intensive firms. There are significant wage growth differentials across ownership categories: the wage growth rates in FIEs and HMTIEs are significantly higher than the wage growth rate in domestic firms. Furthermore, the wage growth in SOEs

Table 5. Wage growth differentials and spillovers: fixed effect model Wage growth

Lagged wage Lagged size Lagged capital intensity Lagged SOEs Lagged COEs Lagged Private firms Lagged HMTIEs Lagged FIEs

Within-industry spillovers

Within-province spillovers

Total (1)

Total (2)

Domestic (3)

Total (4)

Domestic (5)

Total (6)

Domestic (7)

1.576*** (392.9) 0.116*** (58.90) 0.001*** (39.21) 0.087*** (12.36) 0.040*** (8.45) 0.011** (2.51) 0.025*** (2.85) 0.035*** (3.88)

1.616*** (399.56) 0.061*** (30.64) 0.001*** (32.27) 0.074*** (10.58) 0.030*** (6.38) 0.001 (0.25) 0.024*** (2.79) 0.037*** (4.06) 0.022 (0.84) 0.168*** (7.87) Yes 0.42 1,454,584 (385,860)

1.613*** (348.1) 0.055*** (23.37) 0.001*** (23.94) 0.046*** (6.32) 0.009* (1.79) 0.008* (1.84)

1.616*** (399.54) 0.061*** (30.72) 0.001*** (32.26) 0.075*** (10.64) 0.031*** (6.47) 0.001 (0.29) 0.024*** (2.85) 0.036*** (4.04) 0.062*** (3.18) 0.081*** (3.59) Yes 0.42 1,454,584 (385,860)

1.613*** (348.11) 0.055*** (23.48) 0.001*** (23.90) 0.046*** (6.33) 0.009* (1.80) 0.008* (1.94)

1.618*** (399.8) 0.060*** (29.91) 0.001*** (32.12) 0.066*** (9.33) 0.029*** (6.02) 0.002 (0.56) 0.028*** (3.26) 0.041*** (4.60) 0.324*** (4.76) 1.106*** (23.91) Yes 0.43 1,454,584 (385,860)

1.614*** (348.1) 0.054*** (22.95) 0.001*** (23.99) 0.042*** (5.82) 0.010** (2.12) 0.004 (0.88)

Lagged share of HMT employment Lagged share of foreign employment Year dummies Within R2 Observations (Firms)

Within-industry and province spillovers

Yes 0.41 1,454,584 (385,860)

0.112*** (3.32) 0.122*** (4.51) Yes 0.41 1,139,179 (316,882)

Note: The t-values are reported in parentheses. The data source is the enterprise dataset for 1998–2007. * Represent significance at the 10% level. ** Represent significance at the 5% level. *** Represent significance at the 1% levels.

0.073*** (2.97) 0.041 (1.42) Yes 0.41 1,139,179 (316,882)

1.083*** (12.86) 0.635*** (11.51) Yes 0.42 1,139,179 (316,882)

1330

WORLD DEVELOPMENT

and COEs is significantly lower than that in private firms and DJVs. To investigate how FDI presence affects the wage growth in domestic firms, we estimate the wage growth equation augmented by the different measures of FDI presence. Columns 2 and 3 of Table 5 report the results when spillovers are limited to the same industry and province. Columns 4 and 5 report the results when spillovers are limited to the same two-digit industry, and columns 6 and 7 report the results when spillovers are limited to the same province. The estimation results using the full sample are reported in columns 2, 4, and 6 of Table 5, and the results using only the sample of domestic firms are reported in columns 3, 5, and 7. The presence of HMT investment has a significantly negative effect on wage growth in domestic firms. The presence of foreign investment within the same region significantly discourages wage growth in domestic firms, whereas the effect of foreign spillovers within the same industry is insignificant. This pattern is consistent with the results of previous studies of foreign spillover on wage levels. In summary, our study suggests that multinationals not only offer higher wages than domestic firms, but also that their presence lowers the wage level in domestic enterprises and discourages wage growth in those firms. Hence, foreign participation tends to enlarge the inter-enterprise wage gap. 5. CONCLUSION Most developing countries have experienced a sharp increase in inequality over the course of the globalization process. This study presents evidence that exposure to FDI is closely associated with inter-enterprise wage inequality. Using Chinese micro-level data, we investigate foreign wage differentials and foreign spillovers to domestic firms. Consistent with previous studies, we find a significant foreign premium in both wage and non-wage compensation. The presence of foreign and HMT investment has significantly negative spillover effects on the wage level in domestic firms. These negative effects

are more significant for HMT investment than for foreign investment, indicating that the technology externalities of HMTIEs are weaker than those of FIEs. The insignificance of the within-industry spillovers from foreign investment suggests that within-industry technology spillovers may be sufficiently strong to offset the negative spillovers created through pecuniary channels. Our analysis also shows that the wage growth rates in FIEs and HMTIEs are higher than the wage growth rate in domestic firms, and that foreign and HMT investment discourages wage growth in domestic firms. FDI promotion policies are important development strategies for developing countries undergoing industrialization. For example, FDI has been long recognized as an important engine of China’s rapid growth. Whalley and Xin (2006) find that foreign firms (including HMTIEs) contributed more than 40% of China’s economic growth in 2003 and 2004. However, the distribution effect has become an increasingly important consideration in evaluating these policies. Our findings of significant foreign wage differentials and negative foreign wage spillovers suggest that FDI may have a detrimental effect on domestic firms when the negative competition effect dominates the positive effect of technology externalities. Foreign participation not only directly contributes to wage inequality through the composition effect, but also enlarges the wage gap through negative spillovers to domestic enterprises. Due to data limitations, our analysis does not include the service sector. The service sector has gradually been liberalized in recent years, and has become increasingly important to the Chinese economy, accounting for 39.1% of GDP and 32.1% of employment in 2006. Foreign participation in the service industry has grown rapidly, especially since China’s accession to the WTO (see, e.g., Branstetter & Lardy, 2006). It would be interesting for future research to examine the impact of foreign entry to the service sector on wage inequality. Another possible extension of our study would be to investigate whether the wage effect of FDI depends on regional, industry, and enterprise characteristics.

NOTES 1. See Graham (2000) and Brown et al. (2003) for detailed discussions. 2. See the literature survey of Lipsey (2004).

7. The limited sample includes most SOEs and large- and medium- sized non-state enterprises (those with total sales of over 5 million yuan per year).

3. The census covered more than 5 million legal entities in 2004, and the coverage rate was about 97%.

8. The information on skill composition and gender composition is not available in the panel dataset, and thus these variables are not controlled in the panel analysis.

4. State-owned limited liability enterprises are classified as SOEs. Joint ventures between SOEs are treated as SOEs. Joint ventures between COEs are treated as COEs. Domestic cooperative enterprises are treated as COEs.

9. For robustness, we calculate the fixed effect estimations with AR (1), and obtain results that are consistent with those reported in column 1.

5. The results of the mean difference test can be provided upon request. 6. One limitation of this dataset is that it is biased toward large enterprises. This will bias our estimation if foreign wage spillovers to small firms are different from those to large firms.

10. Own-firm employment is excluded in measuring the foreign presence within each region and industry. 11. The information was obtained from the China Association of Development Zones.

REFERENCES Aitken, B., Harrison, A., & Lipsey, R. (1996). Wages and foreign ownership, a comparative study of Mexico, Venezuela, and the United States. Journal of International Economics, 40, 345–371.

Almeida, R. (2007). The labor market effects of foreign owned firms. Journal of International Economics, 72, 75–96.

FOREIGN DIRECT INVESTMENT AND WAGE INEQUALITY Amiti, M., & Javorcki, B. S. (2008). Trade costs and location of foreign firms in China. Journal of Development Economics, 85, 129–149. Bedi, S., & Cieslik, A. (2002). Wages and wage growth in Poland: The role of foreign direct investment. Economics of Transition, 10, 1–27. Blonigen, B., & Ma, A. (2007). Please pass the catch-up: The relative performance of Chinese and foreign firms in Chinese exports. In R. Feenstra, & S. Wei (Eds.), China’s growing role in world trade. The University of Chicago Press. Branstetter, L., & Lardy, N. (2006). China’s embrace of globalization. NBER working paper no. 12373, National Bureau of Economic Research, Cambridge, MA. Braunstein, E., & Brenner, M. (2007). Foreign direct investment and gender wages in urban China. Feminist Economics, 13, 213–237. Brown, D., Deardorff, A., & Stern, R. (2003). The effects of multinational production on wages and working conditions in developing countries. NBER working paper no. 9669. Chen, Y., Demurger, S., & Fournier, M. (2005). Earnings differentials and ownership structure in Chinese enterprises. Economic Development and Cultural Change, 53, 933–958. Cheng, L., & Kwan, Y. (2000). What are the determinants of the location of foreign direct investment? The Chinese experience. Journal of International Economics, 51, 379–400. Conyon, M., Girma, S., Thompson, S., & Wright, P. (2002). The productivity and wage effect of foreign acquisition in the United Kingdom. Journal of Industrial Economics, 50, 85–102. Driffield, N., & Girma, S. (2003). Regional foreign direct investment and wage spillovers: Plant level evidence from the UK electronics industry. Oxford Bulletin of Economics and Statistics, 65, 453–474. Feliciano, Z., & Lipsey, R. (2006). Foreign ownership, wages, and wage changes in U.S. industries, 1987–92. Contemporary Economic Policy, 24, 74–91. Girma, S., Greenaway, D., & Wakelin, K. (2001). Who benefits from foreign direct investment in the UK?. Scottish Journal of Political Economy, 48, 119–133. Girma, S., & Gorg, H. (2007). Evaluating the foreign ownership wage premium using a difference-in-differences matching approach. Journal of International Economics, 72, 97–112. Goldberg, P., & Pavcnik, N. (2007). Distributional effects of globalization in developing countries. Journal of Economic Literature, 45, 39–82. Gorg, H., & Greenaway, D. (2004). Much ado about nothing? Do domestic firms really benefit from foreign direct investment?. World Bank Research Observer, 19, 171–197. Graham, E. (2000). Fighting the wrong enemy: Antiglobal activists and multinational enterprises. Washington, DC: Peterson Institute for International Economics. Gunderson, M. (2006). Viewpoint: Male–female wage differentials: How can that be?. Canadian Journal of Economics, 39, 1–21. Hale, G., & Long, C. (2007). Is there evidence of FDI spillover on Chinese firms’ productivity and innovation? Working paper, Federal Reserve Bank of San Francisco. Head, K., & Ries, J. (1996). Inter-city competition for foreign investment: static and dynamic effects of China’s incentive areas. Journal of Urban Economics, 40, 38–60. Heyman, F., Sjoholm, F., & Tingvall, P. (2007). Is there really a foreign ownership wage premium? Evidence from matched employer–employee data. Journal of International Economics, 73, 355–376. Hu, A., & Jefferson, G. (2002). FDI impact and spillover: Evidence from China’s electronic and textile industries. The World Economy, 25, 1063–1076. Lipsey, R. (1994). Foreign-owned firms and U.S. wages. NBER working paper no. 4691. Lipsey, R. (2004). Home and host effects of FDI. In R. Baldwin, & A. Winters (Eds.), Challenges to globalization (pp. 333–379). Chicago: Chicago University Press. Lipsey, R., & Sjoholm, F. (2002). Foreign firms and Indonesian manufacturing wages: An analysis with panel data. NBER working paper no. 9417. Lipsey, R., & Sjoholm, F. (2004). Foreign direct investment, education and wages in Indonesian manufacturing. Journal of Development Economics, 73, 415–422. Liu, Z. (2008). Foreign direct investment and technology spillovers: Theory and evidence. Journal of Development Economics, 85, 176–193. National Bureau of Statistics of China (1999–2008). China statistical yearbook. Beijing: China Statistics Press.

1331

National Bureau of Statistics of China (1999–2008). China industry economy statistical yearbook. Beijing: China Statistics Press. National Bureau of Statistics in China (2006). China economic census yearbook. Beijing: China Statistics Press. Shen, J., & Deng, X. (2008). Gender wage inequality in the transitional Chinese economy: A critical review of post-reform research. Journal of Organisational Transformation and Social Change, 5, 109–127. Xiao, G. (2005). Round-tripping foreign direct investment and the People’s Republic of China. Tokyo: Asian Development Bank Institute. Wei, Y., & Liu, X. (2006). Productivity spillovers from R&D, exports and FDI in China’s manufacturing sector. Journal of International Business Studies, 37, 544–557. Whalley, J., & Xin, X. (2006). China’s FDI and non-FDI economies and the sustainability of future high Chinese growth. NBER working paper no. 12249.

APPENDIX A A.1 Data sources First National Economic Census conducted in 2004: the basic statistics included in this dataset are summarized in the China Economic Census Yearbook (National Bureau of Statistics in China, 2006). Tables 1 and 2 are based on this dataset. National Bureau of Statistics Enterprise Dataset (1998– 2007): the National Bureau of Statistics of China (NBSC) obtains annual reports from most state enterprises and from large- and medium-sized non-state enterprises (those with sales of more than five million yuan per year). These annual reports contain the firm’s financial statements and some non-financial information, such as the entry date, district code, industry code, and main products of the enterprise. This dataset is used as the basis for compiling basic statistics on the aggregate manufacturing sector that are summarized in the China Statistical Yearbook (NBSC, 1999–2008) and statistics on two-digit manufacturing industries summarized in the China Industry Economy Statistical Yearbook (NBSC, 1999–2008). Tables 3– 5 and Figure 1 are based on this dataset. A.2 Definition of the variables Wage: the logarithm of the average wage level. Ownership dummies: SOE, COE, private enterprise, HMTIE, and FIE. The benchmark is DJVs. Size: the logarithm of total sales. Capital intensity: fixed assets divided by the number of employees. Female share: the proportion female workers of the total number of employees. Skill composition: the proportion of employees with a graduate education (18 years of education and over), the proportion of employees with a college education (16 years of education), and the proportion of employees with a high school education (12 years of education). The benchmark is employees without a high school education. Foreign (HMT) presence in the same region/industry: the share of employment of FIEs and HMTIEs of total employment in the same region/industry. Regional FDI policy index: the number of incentive zones in each province, including Special Economic Zones, economic and technological development zones, and national foreign trade zones.

1332

WORLD DEVELOPMENT

Industry FDI policy index: according to the “Provisional Guidelines for Foreign Investment Projects,” FDI projects in different industries are classified into four categories: encouraged, permitted, restricted, and prohibited. We use three dum-

mies—a permitted dummy, a restricted dummy, and a prohibited dummy—as instruments. The benchmark is an industry in which foreign entry is encouraged.

Available online at www.sciencedirect.com