Export Promotion of OFDI from Emerging Markets Transaction-level Evidence from China Wenjie Chen (GWU, School of Business) Heiwai Tang (Johns Hopkins University, SAIS and CESIfo) GWU China Conference G2 @ GW, Nov 8, 2013 1
Motivation Background in US$ mil 90 000.0 80 000.0 70 000.0 60 000.0 50 000.0 40 000.0 30 000.0 20 000.0 10 000.0 Chinese Outward Direct Investment Flows - China is the world s fifth largest direct investor Headlines: Shuanghui acquires Smithfield CNOOC s plan to invest in Unocal Geely buys Volvo of Sweden Drastic growth of the country s OFDI in recent years deserves further investigation. Source: UNCTAD, 2012 2
Motivation Share of China s OFDI flows 100% 90% 80% Leasing and business services 70% 60% 50% 40% 30% 20% Banking Wholesale and retail trade 10% Mining 0% 2006 2007 2008 2009 2010 Source: China s Ministry of Commerce 3
Motivation Background Typical motives for OFDI: Resource seeking Technology seeking Export promotion (market seeking) The share of OFDI from Leasing and Business Services + Wholesale and Retail Trade accounted for over half of the aggregate volume of China s OFDI flows in 2010 (27 percent in 2006). Benchmark: Mining (8 percent) and banking (13 percent) of OFDI flows in 2010 4
Question Background Common perception is that Chinese OFDI have resource and technology seeking motives Our focus is on export promotion Entering new export market requires overcoming fixed costs, uncertainty, and other business obstacles: Adapting to new customers taste and demand Adjusting to host country s business climate, cultural patterns, structure of market system Complying with host countries local regulations and requirements Solution: Setting up office in host country to tackle obstacles mentioned above 5
Question Research Question Using comprehensive micro data, we ask: Does FDI complement or substitute firm exports? At which margin of exports does FDI complement or substitute exports? How does FDI affect trade performance? Are there third-market effects of FDI on exports? 6
Preview of Using merged data between customs transaction-level trade data and our OFDI firm data (1,800 firms over 2000-2006): Findings: after conducting OFDI, firms Export volume, unit values, varieties, and destinations all increase; Diversification away from core competencies; 7
Preview of con t Findings: after conducting OFDI, firms Capital and skill intensities of exports increases; Third-market effects: increase in unit values, skill and advertising intensity of new products; : FDI from China complements rather than substitutes exports. OFDI also raises level of sophistication of firm s exports It is worth noting that the complementary effects are not confined to the OFDI destination only 8
Literature Review Chinese OFDI: Deng (2003, 2004); Buckley et al. (2007); Antkiewicz and Whalley (2007); Cheng and Ma (2007); Huang and Wang (2011) FDI and exports as substitutes: Brainard (1997); Markusen and Venables (2000); and Helpman, Melitz, and Yeaple (2004) FDI and exports as complements: Lipsey and Weiss (1981, 1984); Yamawaki (2001); Graham (2000); Clausing (2000) Both complementarity and substitution effects: Blonigen (2001) and Conconi et al. (2012) Export-platform production: Antras (2003); Grossman, Helpman, and Szeidl (2003); Yeaple (2003); Conconi et al. 2012 9
Contribution Provides micro evidence on OFDI from developing countries Documents detailed relationship between OFDI and trade at firm and sub-firm level Factor intensities of exports after OFDI New Product Dimension Examines third-market effects 10
OFDI Outward foreign direct investments by Chinese companies provided by the China s Ministry of Commerce. The data set covers OFDI deals between January 1, 1998 and December 31, 2009. Contains information on investing firm's name, firm's industry, province of origin, and the destination of the investment. No value of the deal. 9,744 deals from 7,202 unique firms (most comprehensive we are aware of). Compare the number of deals in our data with those studied by Huang and Wang (2011). Our data set covers 90 percent of the deals from Zhejiang, the province they focus on over the same sample period. 11
OFDI Top 10 OFDI destinations Country Frequency Percent Hong Kong, China 1,946 19.97 United States 918 9.42 Russian FederaBon 551 5.65 Vietnam 464 4.76 United Arab Emirates 370 3.8 Japan 360 3.69 Korea, Rep. 299 3.07 Germany 270 2.77 Lao PDR 267 2.74 Australia 236 2.42
Customs Evidence from Customs Transaction-level Trade
Customs Customs Trade The universe of Chinese import and export transactions 2000 and 2006. on imports and exports (in USD) at the HS 8- digit (>7,000 products) level from a firm to/from each country. 60-80% of the deals can be merged.
Export and OFDI Margin of Exports - Breakdown Follow Bernard, Redding and Schott (2009), decompose firm export volume into various margins: where X i is firm i s volume of exports (in a given year) N ic is # of countries N ih is # of products exported o ich is firm s number of positive country-product transactions A A and 15
Margins of Export Extensive Margin Dependent Var: ln(exp value) Across all Countries ln(# of country) ln(# of HS6) ln(avg Exp) ln(density) OFDI dummy 0.336*** 0.128*** 0.136*** 0.148*** -0.075*** (0.045) (0.020) (0.021) (0.034) (0.014) Firm FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes R-squared.0781.0767.03.0327.0363 Nb of Obs 628048 628048 628048 628048 628048 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. Controlling for firm and year fixed effects
Margins of Export Intensive Margin Dependent Var: Same Country as OFDI Destination ln(exp value) ln(#of HS6) ln(average Exp) OFDI dummy 0.116*** 0.0487*** 0.0674*** (0.014) (0.006) (0.012) Firm FE Yes Yes Yes Year FE Yes Yes Yes R-squared.0322.0144.0225 Nb of Obs 5691703 5691703 5691703 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. Controlling for firm and year fixed effects
OFDI and Customs Unit Value of Products after OFDI Dep Variable: OFDI dummy (OD) 0.187*** (0.012) log Unit Value OD same country 0.214*** (0.029) OD diff country 0.184*** (0.012) OD same; adv 0.269*** (0.049) OD same; dev 0.179*** (0.034) OD diff; adv 0.174*** (0.021) OD diff; dev 0.189*** (0.015) Firm FE Yes Yes Yes Country FE Yes Yes Yes R-squared 0.336 0.336 0.336 Nb of Obs 24,980,644 24,980,644 24,980,644 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. 18
OFDI and Customs Factor Intensity of Products after OFDI Dep Variable: log K int log S int log RD int log ADV int OFDI dummy (OD) 0.187*** 0.006 0.023 0.021 (0.012) (0.004) (0.021) (0.018) Firm FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Nb of Obs 751379 750436 745641 747990 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. Controlling for firm and year fixed effects 19
OFDI and Customs New Product Dimension Ratio of export-weighted average of intensity measure for newly added products to existing products log(k new / K old ) log(k new / K old ) log (S new / S old ) log(s new / S old ) log (RD new / RD old ) log (RD new / RD old ) log (AD new / AD old ) Dep Variable: OFDI Dummy (OD) -0.031+ 0.001-0.01 0.033 log (AD new / AD old ) (0.016) (0.007) (0.040) (0.033) OD same country 0.023 0.026*** 0.065 0.080** (0.017) (0.008) (0.040) (0.034) OD different country 0.014 0.022*** 0.046 0.066** (0.016) (0.008) (0.040) (0.032) Nb of Obs 389610 390740 389203 390332 386044 387166 387881 389008 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. Controlling for firm and year fixed effects 20
OFDI and Customs New Product Dimension Dep Variable: Log(K new /K old ) log(s new /S old ) log(rd new / RD old ) log(ad new / AD old ) OD same; adv -0.001 0.029** 0.054 0.140** (0.027) (0.013) (0.070) (0.056) OD same; dev 0.041+ 0.023** 0.072 0.037 (0.022) (0.010) (0.047) (0.041) OD diff; adv -0.008 0.024+ 0.013 0.124** (0.025) (0.012) (0.066) (0.054) OD diff; dev 0.029 0.021** 0.070 0.022 (0.021) (0.010) (0.048) (0.040) Nb of Obs 390740 390332 387166 389008 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. Controlling for firm and year fixed effects 21
Robustness Checks Exclude firms that invest to Hong Kong Exclude processing firms Different measures of factor intensity 22
Firms conducting OFDI experience increase in # of countries, # of products, avg value of sales per productcountry line, and spread out its sales away from core competencies (i.e., less concentrated). OFDI helps firms export more difficult products (e.g., skill-intensive & advertisement-intensive). We also find interesting third-market effects complementing existing literature Highlight that horizontal OFDI from exporters in developing countries two purposes: Lowers export costs in host countries; Transfer skill and marketing capabilities that help penetrate in other markets as well 23
Thank You! Comments welcome! 24
APPENDIX 25
Examples of OFDI and Export Firms Company: Nanjing Textiles Import & Export Corp Ltd. ( 份 ) Foreign Affiliate: HotPoint Ltd. ( ) Company: Jiangsu Guotai International Group ( 份 ) Foreign Affiliate: 份 Company: Heilongjiang Huayu Industry and Trade Co.,Ltd ( ) Foreign Affiliate : Huahai Wood. Ltd. 宁 ( 宁 ) Company: Xiamen Overseas Chinese Electronic Co., Ltd. ( 份 ) Foreign Affiliate : Sinoprima ( ) Company: Haier Group ( ) Foreign Affiliate : Haier Europe ( ) 26
Characteristics of OFDI Firms from Manufacturing Firm Survey 27
NBS Manufacturing Firm Survey Source: National Bureau of Statistics (NBS),1998-2009 All state-owned firms and all private firms with sales > 5 million RMB (about 600K USD during the sample period) Balanced-sheet variables: firm ownership, output, value added, exports Merge it with our OFDI data based on firm name; 27-55% of the deals can be merged
OFDI and NBS OFDI Effects on Firm Size and Performance Controlling for firm and year fixed effects: Dep Var: ln(sales) ln(value Added) ln(emp) OFDI 2.056*** -0.027 1.809*** (0.342) (0.050) (0.287) R-squared 0.424 0.245 0.369 Nb of Obs. 2352687 1515828 2718572 Dep Var: Labor prod Export/ Sales ln(value Added/ Sales) ln(rd/sales) OFDI 0.278* 0.260*** -0.008 0.298*** (0.108) (0.016) (0.013) (0.078) R-squared 0.56 0.272 0.251 0.187 Nb of Obs. 2531859 2352687 1551447 902316 OFDI = 1 for all firm-years when and after a firm reported overseas investment. The number of observations fluctuates because data for some variables are not available in all years (e.g., R&D only for 2003-2005). Standard errors, clustered at the industry level (2-digit), are in brackets. * p<0.05; ** p<0.01; *** p<0.001
OFDI and NBS OFDI Participation Dependent Var OFDI Dummy Sample All Exclude HK as a Host ln(sales) 0.233*** 0.213*** (0.013) (0.014) ln(empl) 0.0649*** 0.0694*** (0.014) (0.015) SOE - 0.178*** - 0.222*** (0.049) (0.055) HKMT - 0.342*** - 0.409*** (0.027) (0.031) Foreign - 0.289*** - 0.290*** (0.025) (0.026) ln(labor Prod) - 0.009-0.006 (0.011) (0.012) Exp/Sales 0.596*** 0.606*** (0.025) (0.026) Raw material/sales - 0.049-0.070 (0.062) (0.066) Industry FE Yes Yes Province FE Yes Yes N 979884 979544 OFDI = 1 for all firm-years when and after a firm reported overseas investment. The number of observations fluctuates because data for some variables are not available in all years (e.g., R&D only for 2003-2005). Standard errors, clustered at the industry level (2-digit), are in brackets. * p<0.05; ** p<0.01; *** p<0.001
OFDI and Customs Factor Intensity of Products after OFDI Controlling for firm and year fixed effects: Dep Variable: log K int log S int log RD int log ADV int OD same; adv 0.047** 0.038*** -0.009 0.043 (0.018) (0.010) (0.052) (0.045) OD same ; dev -0.011-0.008-0.107*** -0.060** (0.015) (0.007) (0.032) (0.029) OD diff; adv -0.012-0.005 0.023-0.034 (0.015) (0.007) (0.039) (0.033) OD diff ; dev 0.041*** 0.009-0.044-0.021 (0.012) (0.006) (0.030) (0.025) Nb of Obs 752625 751680 746880 749233 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. 31
OFDI and Customs Third Market Effects Controlling for firm and year fixed effects: Dep Variable: log K int log S int log RD int log ADV int Same country 0.016 0.013** -0.063** -0.013 (0.012) (0.006) (0.030) (0.026) Diff country 0.017+ 0.002-0.015-0.027 (0.009) (0.005) (0.024) (0.020) Firm FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Nb of Obs 752625 751680 746880 749233 * p<0.05; ** p<0.01; *** p<0.001. OFDI = 1 for all firm-years when and after a firm reported overseas investment. Robust standard errors reported in brackets. 32